Interdisciplinary Design

Fengfeng Ke • Valerie Shute
Kathleen M. Clark • Gordon Erlebacher
Interdisciplinary
Design of Game-based
Learning Platforms
A Phenomenological Examination
of the Integrative Design of Game,
Learning, and Assessment
Fengfeng Ke Valerie Shute
Department of Educational Psychology Department of Educational Psychology
and Learning Systems and Learning Systems
Florida State University Florida State University
Tallahassee, FL, USA Tallahassee, FL, USA
Kathleen M. Clark Gordon Erlebacher

School of Teacher Education Department of Scientific Computing
Florida State University Florida State University
Tallahassee, FL, USA Tallahassee, FL, USA
ISSN 2567-8086 ISSN 2567-8485 (electronic)

Advances in Game-Based Learning
ISBN 978-3-030-04338-4 ISBN 978-3-030-04339-1 (eBook)
https://doi.org/10.1007/978-3-030-04339-1
Library of Congress Control Number: 2018962489
© Springer Nature Switzerland AG 2019

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, express or implied, with respect to the material contained herein or for any errors
or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims
in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Acknowledgment
The E-Rebuild project is funded by the National Science Foundation, grants

1318784 and 1720533. Any opinions, findings, and conclusions or recommenda-
tions expressed in this book are those of the authors and do not necessarily reflect
the views of the National Science Foundation.
This work would not have been possible without the participation and support of
our E-Rebuild design and research team members. We are especially indebted to
Danial A. Smith, an E-Rebuild lead programmer; Dr. Peyman Fazian, an E-Rebuild
level and assessment programmer; Dr. Anne Taylor, the architecture consultant;
Sungwoong Lee, an E-Rebuild learning support designer; and Biswas Parajuli, an
E-Rebuild assessment programmer. We are also grateful to all teachers and students
who have participated in the design-based research of E-Rebuild.
v
Contents
1 Introduction and Prior Research Review�� 1

2 Chronicle of Designing a Game-Based Learning Platform�� 15
3 Interdisciplinary Design Activities and Patterns�� 51
4 Design of Gameplay for Learning�� 77
5 Interweaving Task Design and In-Game Measurement �� 101
6 Designing Dynamic Support for Game-Based Learning�� 121
7 An Evolving Design Framework for Game-Based
Learning Platforms�� 143
Index�� 155
vii
Chapter 1
Introduction and Prior Research Review
Abstract Games are not just a vehicle to enhance learning but a new way of
understanding and organizing learning. The starting point of an optimal integration
of learning and play in the game setting is to identify the salient elements of learn-
ing itself and the inherent learning processes in gameplay. In this introductory
chapter, we discuss relevant theoretical perspectives on the nature of knowledge
and learning that guide the exploration of playful elements in learning and conse-
quently the opportunities for learning that games offer. We then provide a concep-
tual review of prior research of game design and discuss the challenges and the
interdisciplinary nature of educational game development. In the end, we intro-
duce the overall goal, structure, and chapters of this book.
1.1 Games for Learning
AU1 The best way to motivate students to embrace learning is to “make it fun”, and
hence making learning fun seems the most well-known claim on the educational
power of games (Kim, Song, Lockee, & Burton, 2018). Games are structured or
organized play, requiring the player to follow a specific set of rules to tackle obsta-
cles and attain a goal (Klopfer, Osterweil, & Salen, 2009; Suits, 1978). Although
games are typically considered as a high-quality multisensory rendering environ-
ment, the main reasons for the interest in game for learning are not the medium’s
success but the motivation of players and their deep engagement while playing
(Denis & Jouvelot, 2005).
While learning and play are often associated with opposite connotations
(Mitgutsch, 2009), learning and play share a key attribute: Both are challenging,
long, and interactive processes that require cognitive effort and willingness to
acquire new knowledge or skills. Yet people tend to avoid and dislike challenging
learning experiences in the educational setting while engaging in and enjoy game-
play as a form of hard fun. This observation, along with the interest in creating
intrinsically motivating learning, has triggered the examination of games as a
promising tool and environment for learning. Particularly, prior research has
© Springer Nature Switzerland AG 2019 1

F. Ke et al., Interdisciplinary Design of Game-based Learning Platforms,
Advances in Game-Based Learning,
https://doi.org/10.1007/978-3-030-04339-1_1
2 1 Introduction and Prior Research Review
focused on defining the nature and circumstances of “pleasantly frustrating”

learning experiences and how these are and can be aligned with the gaming expe-
rience (e.g., Gee, 2003; Rieber, 1996). Some scholars have also examined the
phenomena of game engagement from both intrinsic motivation and well-internal-
ized extrinsic motivation perspectives (e.g., Csikszentmihalyi, 1997; Deci &
Ryan, 1985; Malone & Lepper, 1987).
The idea of using games for learning purposes was discussed early by Clark
C. Abt in 1975 and further formulated by other scholars of serious games (e.g.,
Breuer & Bente, 2010; Charsky, 2010; Sawyer, 2003; Susi, Johannesson, &
Backlund, 2007). The educational purpose can be assigned to a game in the design,
or by the context it is used. Games designed primarily for entertainment, like com-
mercial off-the-shelf (COTS) games, can be used in a teaching and learning context,
though its learning effect is an exogenous one.
Despite the innate connection between gameplay and learning, not all forms of
games are equally suitable for learning purposes; simply adding educational mate-
rial to an enjoyable game or presenting it in a game-like setting will not create an
engaging and effective game for learning purposes (Breuer & Bente, 2010; Ke, AU2
2016). A frequent cause of a segmentation between play and learning in a game is
that win criteria for a game do not necessarily equate with the things players are
supposed to learn through playing the game (Van Staalduinen & de Freitas, 2011).
Specifically, there are variant categories of learning in games, including things that
we can learn and are deliberately designed in a game (or intentional learning), things
that we can learn but not necessarily designed into the game (or collateral learning),
and things that we must learn in a game to successfully finish the game (Becker,
2008). The desirable situation is that the first two categories overlap with the third
one, in that both intended and incidental learning are enabled and assured by game-
play. The starting point of an optimal integration of play and learning in the game
setting should be to identify the potential enjoyment (or the playful elements) of
learning itself and the inherent learning processes in gameplay (Rodriguez, 2006).
Hence, games are not just a vehicle to enhance learning but “a new way of under-
standing and organizing learning” (Breuer & Bente, 2010, p. 14).
In this introductory chapter, we describe relevant theoretical perspectives on the
nature of knowledge and learning that help us to explore playful elements in learn-
ing and consequently the opportunities for learning that games offer. We then pro-
vide a review of prior research of game design and discuss the challenges and the
interdisciplinary nature of educational game development. We introduce the overall
goal, structure, and chapters of this book at the end.
1.2 Nature of Knowledge and Learning Pertinent to Games
Dewey defined knowledge as “not information, but a mode of intelligent practice

and habitual disposition of mind” (p. 124, 1910). This definition highlights two
salient elements of knowledge that are apt to frame the design of game-based
1.2 Nature of Knowledge and Learning Pertinent to Games 3
learning: a practice that crafts and enacts intelligence and intelligent habits or
ways of thinking constituted in the practice. The intelligent practice, called reflec-
tive inquiry by Dewey, comprises mainly the processes of identifying a problem
(i.e., a quandary that needs to be resolved), studying the problem through active
engagement, and reaching cognitive conclusions as the problem is resolved. The
central facet is the method by which a cognitive conclusion is reached; hence, the
process/action of identifying, actively studying, and resolving a problem, more
than the cognitive conclusion itself, is the item of value (Dewey, 1929; Hiebert
et al., 1996). During the action of inquiry or problem-solving, scientific attitude—
“willingness to endure a condition of mental unrest and disturbance” or that is
capable of enjoying the doubtful—should be deliberately constructed and is alone
worthy of the title of knowledge (Dewey, 1910, p. 13). In view of that, perfecting
the method of inquiry and constructing scientific attitude are more significant to
realize the meaning of knowledge than that of acquiring and utilizing information
and technical procedures.
Derived from Dewey’s perspective is the argument for problematizing subject
matter knowledge (Hiebert et al., 1996). “Allowing the subject to be problematic
means allowing students to wonder why things are, to inquire, to search for solu-
tions, and to resolve incongruities” (Hiebert et al., 1996, p. 12). The conventional
conception of problem-solving typically focuses on the application of acquired
knowledge in realistic situations, whereas problematizing a topic emphasizes knowl-
edge acquisition. Problematizing mathematics, for example, involves treating expe-
riences as problems, examining or manipulating them via overt doing, and
consequently constructing structural understanding that includes insights into the
structure of mathematical objects, strategies for solving problems (both procedures
used for particular problems and general ways of thought needed to construct the
procedures), and dispositions toward mathematics (e.g., seeing mathematics as an
intellectual activity in which they can participate). Notably, real-life problems pro-
vide a legitimate context for problematizing mathematics, though it is only one
context for it. The value of a real-life problem depends on whether students prob-
lematize the situation and whether it offers the chance for students to acquire struc-
tural understanding after problem-solving. In other words, the knowledge to be
acquired depends as much on the mathematical ideas embedded in the task (e.g., a
game-based inquiry) as on the way it is packaged. Therefore, the focus of the design
research of learning in games is not only on what subject content or information is
integrated but how core game tasks will legitimize subject-problematizing actions to
generate structural understanding.
In alignment with the aforementioned learning perspectives is the proposition to
establish the learning activity or task as an epistemic practice (Eriksson & Lindberg,
2016). With a similar assumption that knowledge is constituted in people’s actions,
an epistemic learning practice is characterized by open and question-generating
objects (or tasks) that the learners work with. Apart from contextualizing a learning
task in relation to students’ everyday interests, it is important to choose and shape a
task so that the students discern that the previous tools and solutions they are famil-
iar with are restricted. When working with the learning task, students get to test
previously developed methods and tools and (re)develop theoretical generalizations

that are characteristic for a specific subject. In other words, the learning task should
be developed to enhance such kinds of theoretical thinking, by requiring that “core
principles” or “conceptual relations” constituting a specific knowing are discerned
and understood through learning actions in a content-rich epistemic practice.
In summary, the above theoretical underpinnings for learning in games suggest
that the acquisition of discipline-specific, structural understanding (or theoretical
thinking), compared with the application or acquisition of static content or informa-
tion, is more crucial for the goal of the learning-play integration. Subsequently,
designing gameplay as an epistemic learning practice that delineates subject-
problematizing actions is a fundamental means to establishing game-based learn-
ing. Educational game design is to create meaningful, interactive, and challenging
experiences involving the user as the conductor of his own intellectual practice and
development.
1.3 Designing Games for Learning
1.3.1 Research of Game Design
The content of a game is its behavior, not the media that streams out of it toward the
player (Hunicke, LeBlanc, & Zubek, 2004). The difference between games and
other design artifacts is that their consumption (by players) is relatively unpredict-
able, because the interaction between players and the coded game subsystems cre-
ates dynamic and often unpredictable behaviors (Hunicke et al., 2004). Gameplay is
hence defined as the structures of player interaction with the game system and inter-
action with other players (Bjork & Holopainen, 2004).
Although game design is a frequent theme of textbooks and articles, a unified
design theory diagnosing and predicting the structural relations between desirable
gameplay and game subsystems is still lacking. The mechanics, dynamics, and aes-
thetics (MDA) framework by Hunicke et al. (2004) is an earlier effort contributed to
this aim and is described as the three sectors of a game system. Mechanics are vari-
ous actions, behaviors, and control mechanisms afforded to the player within a
game context. Dynamics refers to the run-time behavior of the mechanics acting on
player inputs over time or gameplay. Aesthetic represents the desirable emotional
responses or experiences evoked in the player. From the designer’s perspective, the
mechanics give rise to dynamic system behavior, which in turn leads to particular
aesthetic experiences. From the player’s perspective, their respective player or aes-
thetic experiences, such as sensation, fantasy, challenge, fellowship, discovery, and
expression, are brought forth by observable dynamics and, eventually, operable
mechanics. Even though the MDA framework lacks instruction detailing the combi-
nation and proportion of mechanics that will result in variant aesthetic goals, it helps
to describe how and why different games appeal to different players or to the same
1.3 Designing Games for Learning 5
players at different times. It should be noted that in games for learning purposes, the
desirable player experiences are not just emotional responses but more cognitive
involvement. However, inadequate game design research addresses the mechanics
and dynamics in relation to the cognitive facet of player experience.
Extending the MDA framework, Salen and Zimmerman (2005) discussed mean-
ingful play as an emotional and psychological experience in a game that emerges
from the relationship between player action and system outcome. Specifically, they
described that meaningful play occurs when the relationships between gaming
actions and their outcomes in a game are both discernable and integrated into the
larger context of the game. Discernable means that the result of the game action is
communicated to the player in a perceivable way. Integrated means that an action a
player takes not only has immediate significance in the game but also affects the
play experience at a later point in the game. In other words, game design is the pro-
cess by which a designer creates an interactive and engaging system to be encoun-
tered and proactively investigated by a player, from which meaning emerges.
Importantly, interactivity afforded by the coded/designed system comprises not
only functional interactivity (or the usability of the interaction interface) but also
cognitive interactivity—psychological, emotional, and intellectual participation
between a participant and a system.
Prior research of game design tends to depict play experience as the product of
coding operable mechanics and observable interactivity. But there is still a debate
between ludology and narratology in the game design literature. In comparison with
the argument for ludology—systems of puzzles, rules, and interactivity designed to
frame play or fun (e.g., Adams & Dormans, 2012; Koster, 2013; Salen, Tekinbaş, &
Zimmerman, 2004)—the narratology proposition (e.g., Jenkins, 2004) emphasizes
the design of narrative structure as the anchor of play experience, describes the
importance of designing environmental storytelling via spaces and artifacts in a
game, and argues that gameplay enacts a personally meaningful story or life experi-
ence whether story is a defining game feature or not.
Recent discussions of game design (e.g., Adams, 2014; Schell, 2014) manage to
coordinate the perspectives of ludology and narratology by arguing that mechanics
(or interactivity) and environmental storytelling (or embedded narrative) in a game,
jointly and ultimately, serve the purpose of experience design. Play or fun is co-
framed by players’ prior experiences, the designed game system, and the emergent
interactivity (dynamics) between the players and the system. Even though compo-
nents of the experience to be designed are well specified in the game literature, the
paths toward the goal are dynamic and murky. Due to such a dynamic and participa-
tory nature of play experience design, mostly game design is based on the intuition
and experience of designers with the infield experimentation and iterative refining.
Descriptive or analytical analyses of the game elements, design rationales, and
experimented design strategies, however, will help to guide the experimentation.
Particularly, empirical and longitudinal research on variant processes of design cre-
ation and experimentation in diverse game design projects will help us construct the
collective knowledge as well as a unified while scalable theory of game design.
1.3.2 Prior Research of Educational Game Design
Multiple conceptual or design models of educational games, or game-based

learning platforms, have been proposed in the literature of instructional design
and learning technologies. For example, Kiili (2005) proposed an experiential
gaming model that reflects the perspective that learning is a construction of cog-
nitive structures through action or practice in the game world. The model depicts
game challenges (or problems) as the center of game-based learning flow by
initiating active experimentation, reflective observation, and schemata construc-
tion. Kiili’s model outlines the process of experiential game-based learning with
its implications for educational game development but lacks a design description
of the game systems. Amory (2007) described a theoretical framework, called
the game object model, for educational game research and development. The
model sets game space, problem space, and social space as the three components
of the educational game system and prescribes a list of defining features for each
space. For example, drama, role models, interaction, and gestures are listed as
the game space features; explicit knowledge, literacy, and information commu-
nication are considered design elements of the problem space, while the social
space includes computer-mediated communication along with social network
analysis. In general, Amory’s game object model presents a visionary integration
of related theoretical constructs of gaming, learning, and social community. It
has not explained the actual design mechanism and lacked technical definitions
of the theoretical constructs for designers. Similarly, de Freitas and Oliver (2006)
used a four-dimensional framework to summarize the key analytical or evalua-
tion constructs (i.e., learner specifics, pedagogy, representation, and context) of
a virtual-world and gaming-based learning environment. These aforementioned
generic models are all aimed to outline and categorize related constructs under-
lying the integration of gaming and learning processes. They focus on a conjec-
tural or an interpretive analysis of the relationships and structure of these
constructs as a system, from more of an educational researcher perspective.
Other works, such as that of Calvillo-Gámez, Cairns, and Cox (2015) and Dickey
(2007), intend to explore educational game design from the player or user per-
spective. They highlight an interpretive and evaluative analysis of the salient
elements with corresponding design norms of an optimal gaming experience and
then discuss how such design norms can be extended with or transferred to the
instructional design principles.
Empirical studies examining educational game design approaches and fea-
tures, compared with those evaluating a game’s learning effectiveness, are lim-
ited. Our recent review of the literature with the searching phrase of “game
design” + “learning” or “education” generated mostly articles that examined
game design as a learning-by-making activity by participants (e.g., Kafai, 2012;
Robertson & Howells, 2008) or used game design elements in nongame educa-
tional settings (e.g., Deterding, Dixon, Khaled, & Nacke, 2011). Published design
research of educational games is frequently a conjectural or an evaluative analysis
1.3 Designing Games for Learning 7
with existing COTS or serious games (e.g., Annetta, 2010; Dickey, 2007;
Mitgutsch & Alvarado, 2012). Pinelle, Wong, and Stach (2008), for example, dis-
cerned 10 principles of game usability—“the degree to which a player is able to
learn, control, and understand a game”—after reviewing 108 different games
from 6 major game genres (p. 1453). These usability principles, applicable for
game design in general, are allowing users to customize video/audio settings, dif-
ficulty, and game speed, allowing users to skip non-playable and frequently
repeated content, and providing (a) unobstructed views appropriate for the user’s
current actions; (b) intuitive and customizable input mappings; (c) controls that
are easy to manage and that have an appropriate level of sensitivity and respon-
siveness; (d) users with information on game status, providing training and help;
and (e) visual representations that are easy to interpret and that minimize the need
for micromanagement. All these game usability principles are transferrable to the
design practice of educational games and are highly consistent with the infield
experimentation findings of our current project.
On the other hand, there is still a lack of studies, especially design-based ones,
that present infield testing or design experimentation findings governing the
development processes and strategies of educational or serious games (Torbeyns,
Lehtinen, & Elen, 2015; Warren & Jones, 2017). In particular, a rich, data-driven,
and theory-contributing description of educational game design exploration and
findings should help us develop structural and in situ knowledge of game design.
The studies by Andersen, Butler, O’Rourke, Popović and their colleagues (2011,
2014, 2015), for example, examined variant design features and development
mechanisms, from game objective, level progression strategy, and incentive
structure to scaffolding, of a math learning game (called “refraction”). Based on
the play data of online gamers and the perspectives of human computer interac-
tion and educational psychology, their research efforts and findings helped to
illustrate diverse design challenges, the framework of design solution exploration
and generation, and solution validation/refining strategies during the educational
game design process. In another example, Denis and Jouvelot (2005) conducted
a case study on design strategies that were aimed to reconcile learning and fun in
a video game dedicated to music education. Building on the self-determination
motivation theory (Deci & Ryan, 1985), they explored game design patterns or
features that helped to reinforce competence, autonomy, and relatedness in game-
based learning. Their design-based research contributed tangible heuristics of
educational game mechanics design, including reifying values into rules to con-
vey knowledge in interactions rather than static data, providing the player expres-
sive ways to confront with and test rules, tuning usability so that entry barriers
(e.g., technical difficulty or the game’s gender bias) that go against the player’s
urge to practice gameplay will be leveled, sequential embedding of novel chal-
lenges and information in game levels to present a positive slope of the learning
curve, and a natural and stealth assessment mechanism of in-game performance.
All of these previous learning integration game design suggestions have shed
light on our design efforts in this current project and were later corroborated by
our project findings.
1.3.3 E
nduring Challenges and Interdisciplinary Nature
of Game-Based Learning Design
The design of games for learning resembles the innate challenge of entertainment
game design: a definite formulation of gaming behaviors or consequences is diffi-
cult because interactions between players and the coded game system are often
complex, dynamic, and unpredictable. Moreover, play is voluntary and intrinsically
motivating and involves proactive cognitive engagement that allows for the freedom
of effort and interpretation, whereas these characteristics do not automatically apply
to game-based learning (Shute & Ke, 2012). Therefore, existing approaches and
tools for developing entertainment games cannot simply be transferred to educa-
tional games or game-based learning platforms. There are multiple enduring and
distinctive challenges associated with the design of games for learning.
First, problematizing a subject and reinforcing an investigative or epistemic
practice with problem-solving are believed to be crucial to game-based learning, as
discussed above. But not all domain knowledge or topics share the same approach
or advantage in being problematized. Certain learning processes are by nature
abstract and sign-mediated and hence cannot always be situated or contextualized.
Reality isn’t always “fun”—mechanics that impact the dynamics to create the desir-
able emotional responses may negatively impact the game’s ability to recreate the
reality of epistemic practices to be simulated. In particular, we lack understanding
of how to design a game to reinforce purposeful development of abstract or sym-
bolic domain knowledge.
Second, a game is typically an open-ended learning environment that empha-
sizes self-regulated experiential learning as well as reflective inquiry. However, not
all students are by default self-regulated learners or capable of performing sense-
making during game-based problem-solving. Students also differ in their prior
domain knowledge, game skills, motives for gaming, and preferences in relation to
modes of play. Their attention span with textual information in the game world can
be vastly reduced due to the multisensory rendering environment. Given such a
heterogeneous player group and the goal to engage and educate all students with a
single game-based learning platform, designing a motivating while learning-
constructive gaming experience is particularly challenging.
Third, prior research and practice of game-based learning tends to underline a
stealth (or incidental) learning idea. Such a perspective is sensible especially when
only generic scientific attitudes and generic thinking skills are the focus of learning
outcomes. On the other hand, learning by nature is a conscious activity. The desir-
able player responses in game-based learning should not be just emotional but cog-
nitive, metacognitive, and domain-specific. Inclusion of the idea of informed and
intentional learning in games (Annetta, 2010) will entail the integration of gameplay-
coherent learning supports in the game that promote purposeful knowledge con-
struction while maintaining motivation and the game flow. A general game design
framework (e.g., the MDA framework) has to be extended to address these addi-
tional facets of aesthetics or player experience. Corresponding with the inclusion of
1.4 Goal and Structure of the Current Book 9
intentional learning to game-based learning platform design is the challenge to

build and coordinate functional usability/interactivity and cognitive usability/inter-
activity. Rather than creating an intuitive interface, one should also design the inter-
face to be intermediary between gaming actions and action-based knowledge
construction.
Ultimately, another unique change of educational game design is to develop
embedded assessment for game-based learning outcomes that assess performance-
based competency development. Even though recently there is increasing research
on the methods of educational data mining in game-based learning assessment,
designing and using games as both a learning and an assessment tool are challenging
in that gradual learning progression can be in conflict with evidence accumulation
for assessment in the game setting. How to achieve an integral design and develop-
ment of learning tasks and the tasks that discern competency status remains murky.
Interdisciplinary Nature of Game Design Games, acting as an interactive micro-
world, a digital toy, and an environmental storytelling device simultaneously, call
for the integration of multiple disciplinary perspectives and approaches. A game
artifact involves the application of communication/information design, visual
design, human-computer interaction, and software engineering. Designing games as
an effective learning platform, given all of the abovementioned challenges, entails
interdisciplinary knowledge and methods of additional disciplines, such as the sub-
ject domain knowledge, learning system design, psychology, and educational mea-
surement. Interdisciplinary collaboration among the design team members from
these different fields, by itself, is a challenging process. It has to balance and incor-
porate their often conflicting values, ways of thinking, domain-specific language,
and methods, especially those of game designers, instructional designers, content
experts, and psychometricians. Although cross-cultural communication and inter-
disciplinary design are recurrent topics in workplace studies and engineering educa-
tion research, the functional processes and patterns of interdisciplinary design for
game development are generally absent in the literature.
1.4 Goal and Structure of the Current Book
In spite of the plethora of research on game-based learning, the design descriptions

or operational conjectures on how to effectively and intrinsically design intentional,
domain-specific learning in gameplay are scarce. Multiple enduring challenges asso-
ciated with educational game design remain unaddressed. The previous efforts to
explain the framework of game-based learning typically focus on an evaluative anal-
ysis of existing games or present a general and theoretical level of reasoning, thus
lacking tangible guidance for the actual educational game design. Game design is
multidisciplinary in nature, yet few efforts have contributed to describing the inter-
disciplinary design process, techno-pedagogical design knowledge, or accessible
design language.
Based on the 4-year design and research of an architecture simulation game for
mathematical learning (called E-Rebuild), this book is aimed to present a phenom-
enological examination and explanation of a functional design framework for games
in education. It aims to provide a rich description of the experiences and perceptions
of designing a digital learning game and of performing interdisciplinary collabora-
tive design among experts of different fields.
Specifically, it will address both practical and conceptual issues about the design
of play-based learning systems that aligns and interweaves game and learning
mechanics, evidence-centered design of learning tasks and assessment, and in-game
learning support. It also aims to explicate the process of coordinating the interdisci-
plinary language and perspective differences in design communication and negotia-
tion for the future development of learning games.
In Chap. 2, we provide an overview of the integrative, interdisciplinary design
process of E-Rebuild and the adopted phenomenological research approach. Core
phases of designing a game-based learning platform are illustrated with a longitu-
dinal log of the iterative design, experimentation, and refining of E-Rebuild. In
Chap. 2 we also present a positioning statement and design reflection notes of
each co-author (and interdisciplinary design team member). In Chap. 3, we
explain salient interdisciplinary design activities and patterns that emerged from
the E-Rebuild project. The description of the interdisciplinary design patterns in
this chapter consists of a contextualized design narrative/account of core design
processes and an analytical synthesis of core design pattern elements—a design
problem statement with its context and specifics, the solution or technique (with
its key structure or mechanism) to solving the stated problem, and the pattern of
transferring or scaling this design solution or design move. In Chap. 4, we report
design challenges associated with the core components of gameplay and review
the gameplay design propositions and infield test findings of E-Rebuild. We dis-
cuss how domain-specific learning is integrated in and activated by core game
actions, rules, game objects, and the game world design. In Chap. 5, we describe
an integrative-design approach that interweaves game-based task design with in-
game assessment of learning. We discuss core design processes and functional
conjectures on the approaches of task generation and evidence accumulation, with
support of infield observations on the implementation feasibility and outcomes of
the design assumptions. In Chap. 6, we explore the design of dynamic game-
based learning support. We review prevalent practices and prior research of scaf-
folding and support in game-based learning, share our observations of the
obstacles that learners experienced in game-based learning processes, and report
the corresponding learning support strategies with infield testing findings of
E-Rebuild. In the conclusive Chap. 7, we discuss an emerged, functional design
framework for game-based learning platforms, with a summary of the design
problem structuring process for the interdisciplinary, integrative design of game-
based learning. We conclude the chapter and the book by discussing potential
directions of future design and research efforts of game-based learning design,
with a set of core design concepts highlighted.
References 11
AU3 References
Abt, C. C. (1975). Serious games. New York: Viking Compass.

Adams, E. (2014). Fundamentals of game design. New York: Pearson Education.
Adams, E., & Dormans, J. (2012). Game mechanics: Advanced game design. Berkeley, CA: New
Riders.
Amory, A. (2007). Game object model version II: A theoretical framework for educational game
development. Educational Technology Research and Development, 55(1), 51–77.
Andersen, E., Liu, Y. E., Snider, R., Szeto, R., Cooper, S., & Popović, Z. (2011, June). On the
harmfulness of secondary game objectives. In Proceedings of the 6th International Conference
on Foundations of Digital Games (pp. 30–37). New York: ACM.
Annetta, L. A. (2010). The “I's” have it: A framework for serious educational game design. Review
of General Psychology, 14(2), 105–112.
Becker, K. (2008). The invention of good games: Understanding learning design in commercial
video games (Doctoral dissertation, University of Calgary).
Breuer, J., & Bente, G. (2010). Why so serious? On the relation of serious games and learning.
Journal for Computer Game Culture, 4(1), 7–24.
Butler, E., Andersen, E., Smith, A. M., Gulwani, S., & Popović, Z. (2015, April). Automatic game
progression design through analysis of solution features. In Proceedings of the 33rd annual
ACM conference on human factors in computing systems (pp. 2407–2416). New York: ACM.
Calvillo-Gámez, E. H., Cairns, P., & Cox, A. L. (2015). Assessing the core elements of the gam-
ing experience. In Game user experience evaluation (pp. 37–62). New York: Springer, Cham.
Cetina, K. K. (2007). Culture in global knowledge societies: Knowledge cultures and epistemic
cultures. Interdisciplinary Science Reviews, 32(4), 361–375.
Charsky, D. (2010). From edutainment to serious games: A change in the use of game characteris-
tics. Games and culture, 5(2), 177–198.
Csikszentmihalyi, M. (1997). Finding flow: The psychology of engagement with everyday life.
New York: Basic Books.
Deci, E., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human behavior.
New York: Springer Science & Business Media.
De Freitas, S., & Oliver, M. (2006). How can exploratory learning with games and simulations
within the curriculum be most effectively evaluated? Computers & Education, 46(3), 249–264.
Denis, G., & Jouvelot, P. (2005, June). Motivation-driven educational game design: Applying
best practices to music education. In Proceedings of the 2005 ACM SIGCHI International
Conference on Advances in computer entertainment technology (pp. 462–465). New York:
ACM.
Deterding, S., Dixon, D., Khaled, R., & Nacke, L. (2011, September). From game design elements
to gamefulness: Defining gamification. In Proceedings of the 15th international academic
AU4 MindTrek conference: Envisioning future media environments (pp. 9–15). ACM.
Dewey, J. (1910). Science as subject-matter and as method. Science, 31, 121–127.
Dewey, J. (1929). The quest of certainty: A study of the relation of knowledge and action.
New York: G.P. Putnam’s Sons.
Dickey, M. D. (2007). Game design and learning: A conjectural analysis of how massively multiple
online role-playing games (MMORPGs) foster intrinsic motivation. Educational Technology
Research and Development, 55(3), 253–273.
Eriksson, I., & Lindberg, V. (2016). Enriching ‘learning activity’ with ‘epistemic practices’–
enhancing students’ epistemic agency and authority. Nordic Journal of Studies in Educational
Policy, 2016(1), 32432.
Gee, J. P. (2003). What video games have to teach us about learning and literacy. Computers in
Entertainment (CIE), 1(1), 20–20.
Hiebert, J., Carpenter, T. P., Fennema, E., Fuson, K., Human, P., Murray, H., et al. (1996).
Problem solving as a basis for reform in curriculum and instruction: The case of mathematics.
Educational Researcher, 25(4), 12–21.
Hunicke, R., LeBlanc, M., & Zubek, R. (2004, July). MDA: A formal approach to game design
and game research. In Proceedings of the AAAI workshop on challenges in game AI (Vol. 4(1),
p. 1722–1726). AU5
Jenkins, H. (2004). Game design as narrative architecture. In P. Harrington & N. Frup-Waldrop
(Eds.), First person (pp. 44–53). Cambridge, MA: MIT Press.
Kafai, Y. B. (2012). Minds in play: Computer game design as a context for children’s learning.
New York: Routledge.
Kiili, K. (2005). Digital game-based learning: Towards an experiential gaming model. The Internet
and Higher Education, 8(1), 13–24.
Kim, S., Song, K., Lockee, B., & Burton, J. (2018). Engagement and fun. In Gamification in
Learning and Education (pp. 7–14). Cham, Switzerland: Springer.
Klopfer, E., Osterweil, S., & Salen, K. (2009). Moving learning games forward. Cambridge, MA:
The Education Arcade.
Koster, R. (2013). Theory of fun for game design. Sebastopol, CA: O’Reilly Media.
Malone, T. W., & Lepper, M. R. (1987). Making learning fun: A taxonomy of intrinsic motivations
for learning. In R. E. Snow & M. J. Farr (Eds.), Aptitude, learning, and instruction volume
3: Conative and affective process analyses (pp. 223–253). Hillsdale, NJ: Lawrence Erlbaum
Associates.
Mislevy, R. J., Almond, R. G., & Lukas, J. F. (2003). A brief introduction to evidence-centered
design. ETS Research Report Series, 2003(1). AU6
Mitgutsch, K. (2009). Passionate digital play-based Learning, (re)Learning in computer games like
Shadow of the Colossus. Eludamos, 3(1), 9–22.
Mitgutsch, K., & Alvarado, N. (2012, May). Purposeful by design?: A serious game design assess-
ment framework. In Proceedings of the International Conference on the foundations of digital
games (pp. 121–128). New York: ACM.
O'Rourke, E., Andersen, E., Gulwani, S., & Popović, Z. (2015, April). A framework for automati-
cally generating interactive instructional scaffolding. In Proceedings of the 33rd annual ACM
conference on human factors in computing systems (pp. 1545–1554). New York: ACM.
O'Rourke, E., Haimovitz, K., Ballweber, C., Dweck, C., & Popović, Z. (2014, April). Brain
points: A growth mindset incentive structure boosts persistence in an educational game. In
Proceedings of the SIGCHI conference on human factors in computing systems (pp. 3339–
3348). New York: ACM.
Pinelle, D., Wong, N., & Stach, T. (2008, April). Heuristic evaluation for games: Usability prin-
ciples for video game design. In Proceedings of the SIGCHI conference on human factors in
computing systems (pp. 1453–1462). New York: ACM.
Rieber, L. P. (1996). Seriously considering play: Designing interactive learning environments
based on the blending of microworlds, simulations, and games. Educational Technology
Research & Development, 44(2), 43–58.
Robertson, J., & Howells, C. (2008). Computer game design: Opportunities for successful learn-
ing. Computers & Education, 50(2), 559–578.
Rodriguez, H. (2006). The playful and the serious: An approximation to Huizinga’s Homo Ludens.
Game Studies, 6(1), 1604–7982.
Salen, K., Tekinbaş, K. S., & Zimmerman, E. (2004). Rules of play: Game design fundamentals.
Boston: MIT press.
Salen, K., & Zimmerman, E. (2005). Game design and meaningful play. In J. Raessens &
J. Goldstein (Eds.), Handbook of computer game studies (pp. 59–79). Cambridge, MA: MIT
Press.
Sawyer, B. (2003) Serious games: Improving public policy through game-based learning and sim-
ulation. Woodrow Wilson International Center for Scholars. ??? AU7
Schell, J. (2014). The art of game design: A book of lenses. New York: AK Peters/CRC Press.
References 13
Shute, V. J., & Ke, F. (2012). Games, learning, and assessment. In D. Ifenthaler, D. Eseryel, &
X. Ge (Eds.), Assessment in game-based learning: Foundations, innovations, and perspectives
(pp. 43–58). New York: Springer.
Suits, B. H. (1978). The grasshopper: Games, life and utopia. Toronto, ON: University of Toronto
Press.
Susi, T., Johannesson, M., & Backlund, P. (2007). Serious games: An overview. Retrieved from:
https://www.diva-portal.org/smash/get/diva2:2416/FULLTEXT01.pdf
Van Staalduinen, J. P., & de Freitas, S. (2011). A game-based learning framework: Linking game
design and learning. In M. S. Khine (Ed.), Learning to play: Exploring the future of education
with video games (pp. 29–53). New York: Peter Lang.
Warren, S. J., & Jones, G. (2017). Learning games: The science and art of development. Cham,
Switzerland: Springer.
Author Queries
Chapter No.: 1 427373_1_En_1_Chapter
Queries Details Required Author’s Response

AU1 Please confirm the identified head levels are okay.
AU2 References “Ke (2016), Bjork & Holopainen (2004), Torbeyns,
Lehtinen, & Elen, (2015), Andersen et al. (2014, 2015)” have
been cited in the text but not provided in the reference list.
Please check.
AU3 References “Butler et al. (2015), Cetina (2007), Mislevy et al.
(2003), O’Rourke et al. (2015), O’Rourke et al. (2014)” were
not cited anywhere in the text. Please provide in-text citation
or delete the reference from the reference list.
AU4 Please provide publisher location (City) for Deterding et al.
(2011).
AU5 Please provide publisher name and location for Hunicke et al.
(2004).
AU6 Please provide page range for Mislevy et al. (2003).
AU7 Please provide complete details for Sawyer (2003).
Chapter 2 1
Chronicle of Designing a Game-Based 2
Learning Platform 3
Abstract Our phenomenological examination of learning game design is situated 4

in a 4-year, longitudinal design-based research project that encompasses iterative 5
design processes to develop, refine, and study a game-based learning platform 6
called E-Rebuild. This chapter presents an introductory overview of the four facets 7
of the interdisciplinary educational game design—interdisciplinary collaboration, 8
learning-play integration, integrative task and assessment design, and game-based 9
learning support. We then provide a design chronicle of E-Rebuild as the key setting 10
of the phenomenon examined, by explaining its iterative design, testing, and refin- 11
ing processes. The authors’ researcher positionality and reflective summaries of 12
design experiences are presented as well. 13
2.1 Introduction 14
AU1 Adopting a phenomenological perspective, we aim to describe as rich and accu- 15

rately as possible the phenomenon of designing a game for the learning purpose (or 16
a game-based learning platform), from the perspectives of people involved while 17
refraining from pre-given frameworks (Giorgi, 2009; Groenewald, 2004). The 18
examination evolves around two overarching questions: What is it like to design a 19
game for the learning purpose? What salient design patterns, strategies, and under- 20
standings have been derived from the design effort? 21
Our phenomenological examination is situated in a 4-year, longitudinal design- 22
based research project that encompasses iterative design processes to develop, refine, 23
and study the game-based learning platform in situ. Specifically, we have been 24
designing and developing an architecture simulation game-based mathematics learn- 25
ing platform (called Earthquake Rebuild or E-Rebuild) that also incorporates evi- 26
dence-centered learning assessment via an unobtrusive collection and analysis of 27
performance data. A major conjecture of E-Rebuild is that architectural design and 28
modeling via this simulation game-based learning platform will engage middle- 29
school students in problem-centered learning and epistemic practice of 30
mathematics. 31

https://doi.org/10.1007/978-3-030-04339-1_2
16 2 Chronicle of Designing a Game-Based Learning Platform
32 Informants or participants are nine members of the design team of E-Rebuild.

33 They include the four authors of this book, in addition to one learning scientist,
34 two game programmers who are graduate students who majored (or are majoring)
35 in scientific computing and computer science, one student instructional designer,
36 and one architect consultant. We have examined the data derived from archived
37 design and meeting notes, open-ended interviews, observation notes, and an arti-
38 fact analysis with the game that was iteratively designed and refined during the 3
39 project years.
40 We analyzed the qualitative data in an attempt to identify themes and make func-
41 tional generalizations regarding the nature and perceived experience of collabora-
42 tively designing a game-based learning platform by an interdisciplinary team.
43 Following a general framework of phenomenological inquiry (Giorgi, 2009;
44 Groenewald, 2004), we have bracketed our positionality (as described in a later sec-
45 tion), delineated and clustered units of meaning from the data to form the themes,
46 and then extracted general and unique themes to make a composite description of
47 every facet of the phenomenon. These thematic descriptions of the interdisciplinary
48 game design are presented in Chaps. 3, 4, 5, and 6.
49 In this chapter, we present an introductory overview of the four facets of the
50 interdisciplinary educational game design that will be discussed in details in later
51 chapters. We then provide a design chronicle of E-Rebuild as the key setting of the
52 phenomenon examined, by explaining its iterative design, testing, and refining pro-
53 cesses. Ultimately, the authors’ researcher positionality and reflective summaries of
54 design experiences are presented.
55 2.2 C
ore Facets of Interdisciplinary Educational Game
56 Design
57 Four core facets of game design, governing the heuristics of interdisciplinary col-
58 laboration, learning-play integration, integrative task and assessment design, and
59 game-based learning support, have emerged from the thematic analysis of quali-
60 tative data. The first facet depicts general patterns framing the allocation and
61 negotiation of collaborative design efforts during every phase of game design.
62 The second facet sets the tone for the rest of game design with a specification of
63 core gameplay —what, in which way, and when learning is legitimized by basic
64 game actions, rules, and settings. The third facet involves a systematic develop-
65 ment of game tasks that exemplify and assemble core actions in meaningful con-
66 texts, to not only enable subject-specific learning progressions but also discern
67 and accumulate performance evidence of competency development. The fourth
68 facet captures the design endeavors to reinforce players’ motivated and inten-
69 tional cognitive efforts contributed to game action-based knowledge application
70 and acquisition.
2.2 Core Facets of Interdisciplinary Educational Game Design 17
2.2.1 Managing Interdisciplinary Collaborative Design 71
The challenges of collaborative design by an interdisciplinary team relate to the 72

management of unshared languages, values, and habitual methods ascribed to dif- 73
ferent forms of design, technological, and educational endeavor, within which is 74
a mutual lack of awareness of each other’s theoretical stance and motives about 75
game-based learning. As Weingart (2000) identified, the way in which we see the 76
world and approach novel problems is affected by a host of subcultures that are 77
specific to individual disciplines, making truly interdisciplinary working difficult 78
to achieve. The disciplines of education, mathematics, architecture, and scientific 79
computing, in this case, have their own theoretical or applied stance toward edu- 80
cation or design, criteria for appraising quality (e.g., of an effective learning 81
experience or a design artifact), and their own ways of working (e.g., with design 82
and teamwork). 83
For example, in spite of a general consensus among the interdisciplinary 84
design team on the importance of engaging students in authentic and constructive 85
learning, we differed in our original understanding of the underlying design con- 86
structs, such as the definition of a game, the nature of mathematics, the dynamics 87
between learning and play (or fun), and interdisciplinary subject matters (e.g., 88
mathematics in architecture). These different theoretical stances are exacerbated 89
by our differences in epistemic frames (e.g., being designer, programmer, educa- 90
tor, or evaluator), experience of gaming technologies as both a consumer and a 91
developer, understanding of the target users (e.g., middle-school students and the 92
school setting), as well as our prioritized motives (e.g., research of game design 93
or learning assessment, mathematics education, or design education). All these 94
differences are reflected in language, design ideas or concepts, and preferred rep- 95
resentation of concepts. 96
Given such a heterogeneity, the interdisciplinary collaborative design is never 97
a simply division of work or a sum of ideas or expertise. Instead, there is an 98
intentional and ongoing effort contributed to building a common design culture 99
among the interdisciplinary team, including the specification of common con- 100
cepts and language, the explanation and selection of the theoretical and design 101
conjectures underlying the aforementioned design constructs, and the negotiation 102
of an overarching design goal with a tangible design agenda that reflects the 103
dynamics of different epistemic perspectives on learning, game, assessment, and 104
the embedded subject matters. Last but not least, the team has explored and grad- 105
ually established an informal protocol or routine for collaborative working and 106
decision-making. This protocol deals with the balance between collaborative 107
brainstorming and efficient decision-making, the integration of self-governing 108
obligations and overlapped responsibilities, and the approach of information 109
sharing and archiving. 110
111 2.2.2 Designing Gameplay for Learning
112 Integrating the design perspective about an optimal integration of play and learning
113 in gaming (e.g., Breuer & Bente, 2010; Rodriguez, 2006) and the argument for
114 constituting knowledge in actions (Dewey, 1929; Hiebert et al., 1996), coding game
115 mechanics (or interactivity) evolves around the effort of understanding and organiz-
116 ing mathematics as inquiries or subject-problematizing practices. As the data illus-
117 trated, the exploration of core gameplay involves two dimensions: (a) specifying a
118 series of game actions that embody mathematical learning as an intellectual inquiry
119 or practice and (b) designing coherent game objectives, obstacles (constraints), and
120 rewarding rules that legitimize an active performance of these actions by the play-
121 ers, thus creating the desirable, learning-constructive interactivity between the play-
122 ers and the game system.
123 Along with the core game action and rule set are the interactivity interface (or the
124 player input/output controls) and the game settings (or environmental storytelling)
125 that should work coherently to foster and motivate emergent functional and cogni-
126 tive interactivity between the players and the game system. In a game-based learn-
127 ing platform, a user interface should not only promote an “intuitive” action but also
128 intermediary, discipline-specific theoretical thinking during the action. Designing
129 game settings (or a game world) for the learning purpose is more than environmen-
130 tal storytelling. Game objects and scenarios in E-Rebuild, for example, embody the
131 external representation of game-based, interactive mathematical problems for the
132 players to investigate and solve. Therefore, designing gameplay for learning is to
133 design a series of subject-problematizing actions that are structured by a set of rules
134 and obstacle-tackling objectives, an intermediary action interface, and a coherent
135 game world as both the milieu and drive of the actions.
136 2.2.3 I ntegrative Design of Game-Based Learning Task

137 and Assessment
138 A fundamental theoretical stance toward the nature of game-based learning (Eriksson
139 & Lindberg, 2016; Hiebert et al., 1996), as discussed in Chap. 1, is that the knowl-
140 edge to be acquired depends as much on the mathematical ideas embedded in the
141 task as on the way it is packaged. Therefore, the focus of the design is not only on
142 what subject content or information is integrated but how core game tasks will exem-
143 plify subject-problematizing actions to generate understanding. In other words, the
144 game-based learning task should be developed to enhance sign-mediated, subject-
145 related theoretical thinking (Eriksson & Lindberg, 2016), by requiring that “core
146 principles” or “conceptual relations” constituting a specific knowing are discerned
147 and understood through learning actions in a game-based task.
2.2 Core Facets of Interdisciplinary Educational Game Design 19
Apart from designing inquiry- and understanding-generating tasks, another core 148
design challenge related to intentional learning in gaming is stealth assessment that 149
enables the real-time capture and analysis of performance-based competency devel- 150
opment data. With the recent methodology development in educational data mining 151
and learning analytics, stealth learning assessment based on the dynamic perfor- 152
mance of players could better capture process and performance-oriented evidence 153
on competency development while not being intrusive to distract players’ gameplay 154
or state of flow. 155
Using the evidence-centered design approach (Almond, Mislevy, Steinberg, Yan, 156
& Williamson, 2015; Mislevy & Haertel, 2006), the project of E-Rebuild will illus- 157
trate an integrative design approach that intertwines learning task development with 158
assessment construction. The selection and construction of game-based learning 159
tasks will elicit the core gaming/learning actions that operationalize the targeted 160
competencies and serve as both the source and evidence of competency develop- 161
ment. The game log is developed to capture these actions, with the logged data 162
parsed as the values of observables (variables) in a competency-based statistical 163
model (e.g., Bayesian networks). The sequencing and accumulation of a collection 164
of tasks (or game levels) then need to support both learning progression and the 165
accumulation of evidence for stealth assessment. 166
2.2.4 Designing Game-Based Learning Supports 167
A main challenge of designing games is that their consumption (by players) is rela- 168
tively unpredictable (Hunicke, LeBlanc, & Zubek, 2004). For an educational game, 169
the interaction between the player and the coded game system is even more dynamic 170
due to the anticipated learning interactivity. Differing in their prior competency 171
status, learning and gaming preferences, and dispositions toward the subject matter 172
and gameplay, players will differ in their reactions and behaviors in a game-based 173
learning system. Apart from learning game mechanics and user interface, the play- 174
ers may be involved in shortcut solutions rather than expected learning actions 175
(Butler, Andersen, Smith, Gulwani, & Popović, 2015), or they may fail to be purpo- 176
sive or mindful to generate theoretical thinking from game tasks and actions. 177
Therefore, support for game-based learning should be designed to not only 178
reduce the entry barriers (e.g., via training) but also enhance the opportunities of 179
intentional learning and knowledge acquisition during gaming for diverse learn- 180
ers. Game-based learning support can be tool- or material-mediated support 181
embedded in the game or external, human agent-mediated support arranged as 182
part of the game-based learning environment. The design efforts focus on explor- 183
ing what, when, and how learning supports are integrated and presented during 184
gameplay to foster task and action engagement and performance for knowledge 185
acquisition. 186
187 2.3 C
hronicle of Iterative Design, Testing, and Refinement
188 of E-Rebuild
189 E-Rebuild creation is a heavily iterative design process. There are more than 100
190 published versions of the game, with either major or minor revisions in between
191 these versions. Below we present a design chronicle recording major developments
192 and revisions associated with E-Rebuild. Iterative player testing was conducted dur-
193 ing and across the following design phases.
194 1. Starting exploring the action of building: In the original version of a build, drag-
195 ging objects around with the mouse in a top-down view was the only control
196 mechanism. The objective was to fill a certain percentage of the given space with
197 shipping containers. The containers were little more than rectangular prisms.
198 After a given time limit, the space would shake to simulate an earthquake. If the
199 same amount of containers remained, the level was passed. The objectives were
200 to quickly simulate an earthquake, track occupied space, and enable container
201 manipulation in a three-dimensional space.
202 2. Designing interface of site surveying and collection: In our next iteration, we
203 decided a fixed view offered limited flexibility and experimented with a first-
204 person view that would offer the freedom to explore the world and give the
205 player a sense of presence. Moving items would happen by picking them up. The
206 player would approach an object, press a key, and carry the item with them. We
207 also introduced a “stamina” meter to stop the user from carrying items forever.
208 This was a great improvement from the previous version but not without
209 compromise. Where the previous iteration required only mouse movement and a
210 single click, the new control scheme was inspired by the WASD control scheme.
211 This added a total of six keys and constant mouse control. For experienced gam-
212 ers, this may be second nature, but experienced gamers do not comprise the
213 totality of our target demographic (Fig. 2.1).
214 3. Designing alternative modes of gameplay, controls for building, and initial game
215 world: After switching to keyboard and mouse movement, we doubled down on
216 the controls. Items could be held from a distance, and then they moved following
217 the player at the distance they were picked up. Using another pair of keys, the
218 player could move the object closer to or farther from the player. An earthquake
219 was simulated on the platform shown in the mini-map. To complete the level,
220 players should move a certain number of containers to the platform and maintain
221 a given height after the earthquake (Fig. 2.2).
222 A toolbar was added to control many options. A number of tools were available.
223 These included marking tools (such as freehand, line, rectangle, and ellipse) and
224 shaping tools, (such as cutting and scaling). We began working on the beginnings of
225 the building mode of E-Rebuild. The view was static with an unmoveable camera.
226 This became an issue when working in three dimensions (Fig. 2.3).
227 At the same time, we started to explore game world design by experimenting
228 with larger environments that appeared to overwhelm players with the opportunity
2.3 Chronicle of Iterative Design, Testing, and Refinement of E-Rebuild 21
Fig. 2.1 Earliest prototype with the basic item-movement control and a stamina meter
Fig. 2.2 Alternative perspective of play and basic controls

Fig. 2.3 A tool bar was added
229 to explore. We settled on a somewhat middle ground—a smaller island. This island
230 remained the basis for the island episode for a number of iterations, which was also
231 the beginning set of the adventure mode of E-Rebuild. In the adventure mode, the
232 player was to explore the area and collect items via a magic wand that made a green
233 “poof” as the item was collected and subsequently disappeared. Collected contain-
234 ers were marked on the bottom-left corner, while additional resources collected
235 were given as credits on the bottom-right corner (Figs. 2.4 and 2.5).
236 4. Designing the allocation action, creating the initial game level with the reward
237 mechanism: A family allocation objective was added. Blue capsules were used
238 as placeholders for people (families). The island people had three states: lost,
239 found, and assigned. Collecting and allocating families into the shelter increased
240 happiness credits, which gave the game three classes of credits: time, resource,
241 and happiness. These can be seen on the bottom right (Fig. 2.6).
242 5. Refining the design of the initial game level: To enable (family) allocation
243 inside a structure (e.g., a container), we refined the collection interface so that
244 players could switch their wand from collecting a whole structure to only col-
245 lect its parts (e.g., panels). This also allowed the player to make windows and
246 entryways in the structure. The building and adventure modes were integrated
247 in a single level: Players could toggle the mode by pressing 0 (zero). In the
248 initial level, the player would collect and place containers to rebuild a row
249 house and then assign the families into the house (Fig. 2.7).
Figs. 2.4 and 2.5 Initial island for the adventure mode of gameplay
6. Developing additional in-game learning supports: An in-game scratchpad was 250

designed to allow users to make quick notes and conduct calculations. Two 251
dimensions (2D) of the target structure (e.g., the row house) and site landmarks 252
were added to assist the player in building the structure. An issue remained 253
unaddressed: it was difficult for the players to get out of the water when they 254
were distracted and explored the underwater land (Figs. 2.8 and 2.9). 255
Fig. 2.6 The first game level
Fig. 2.7 Adding an allocation action

Fig. 2.8 Adding in-game learning supports
Fig. 2.9 Getting stuck under the water

Figs. 2.10 and 2.11 Chunking the task
256 7. Chunking the task in the game level: The task is a complex multistep problem
257 that has many steps that could stop the player from progressing. We tried chunk-
258 ing the task or game level into sub-goals to guide the player, with a progress bar
259 showing the distance to the goal. A tutorial level was also created to train the
260 player on game controls (Figs. 2.10, 2.11, and 2.12). AU2
AU3
261 8. Designing another game level featuring the building action at a higher granular-
262 ity degree and a trading action: With an initial in-field testing study completed,
263 we began to design the desert level that features the building action at a higher
264 granularity and complexity level (e.g., via adobe blocks rather than containers)
265 and a different scene/setting (i.e., a desert). We also added a training action (in
Fig. 2.12 An initial tutorial level
addition to the collection action) with the building materials. Considering the 266
game world exploration issue, we designed natural boundaries such as a sand- 267
storm, a fence, and tall bluffs to the desert set to keep the players on the con- 268
struction site without wandering away (Figs. 2.13 and 2.14). 269
9. Refining the interface of the trading and building actions: We refined the user 270
input interface of the trading action to necessitate cost planning and calculation 271
by the player. In the original movement control of the building action, the item 272
movement followed the cursor, but items would easily overlap with each other 273
and cause a popcorn effect. To fix this issue, we changed the movement control 274
so that the player moved the objects along the two displayed axes that can be 275
changed by moving the camera around the object. Items actually moved on a 276
variable-sized grid. The actual location would be with the green outline when 277
the player decided to place the item (Figs. 2.15 and 2.16). 278
10. Expanding the “desert” and “island” episodes with levels (tasks) development: 279
Both episodes were expanded with tasks (levels) development. As more levels 280
became available, an episode tab became more useful. The episode tab of the 281
menu gave a wide overview of the tasks for this landscape. The task tab gave 282
specific instructions for completing each level. As the levels and game grew more 283
complex and longer in duration, a way to save was necessary. Originally the 284
saves were all local. This forced the user to use the same computer from session 285
to session. Levels with variant difficulty options were given to the user. This dif- 286
ficulty selection was associated with variables and their dependencies in a task, 287
the number of problem-solving steps involved, and the complexity of the embed- 288
ded mathematical competency. To assist with problem-solving, an interactive 289
problem-solving help section was added. Levels were also even more discreetly 290
chunked with more frequent performance feedback (Figs. 2.17, 2.18, and 2.19). 291
Figs. 2.13 and 2.14 Adding a “desert” level featuring building at a higher granularity level
292 11. Adding a “school” episode and more levels: A school set was added as an urban
293 environment in which complicated building tasks, such as classroom and sta-
294 dium stands building, were developed (Figs. 2.20 and 2.21).
295 12. Creating cohesive UI: A new UI was created to be cohesive across the game
296 episodes. The new version opted to use the level chunks as smaller individual
297 levels, each with a single task to complete with a number of constraints (or
298 obstacles). This version also saw the immersion of adventure and building
299 modes—the player does not need to shift between the two modes through a
300 toggle but has a “fly” mode that fully integrates ego- and exocentric perspec-
Figs. 2.15 and 2.16 Trading action and the item movement control
tives in the virtual game world. The interface of the collection action was 301
updated to require action-related mathematical reasoning. The end-of-level, 302
textual performance review was replaced with a visual, star system. The player 303
profiles were associated with game user names to help organize their gaming 304
logs (Figs. 2.22, 2.23, 2.24, and 2.25). 305
13. Creating a cohesive game world and a new “farm” game episode: The level set 306
design was updated to be consistent across game episodes. More levels/tasks 307
were added throughout the game, including a “farm” episode, to better support 308
the learning and assessment of mathematical competencies specified (e.g., ratio 309
Fig. 2.17 “Desert” being a multi-level game episode
Fig. 2.18 Interactive problem-solving help

Fig. 2.19 End-of-level textual performance feedback
Fig. 2.20 Adding a school episode

Fig. 2.21 Being a 3-episode and 13-level game
Fig. 2.22 A new cohesive UI

Fig. 2.23 Intermediary interface of the collection action
Fig. 2.24 A visual, 3-star system as the end-of-level performance feedback

Fig. 2.25 Player profile associated with a game user name for log organization
Fig. 2.26 A cohesive game world: Island
310 and proportion, composition and decomposition of shapes, and transformation

311 of geometric shapes, angle, perimeter, area, volume) while presenting a gradual
312 learning progression. At this point, the game had 19 levels in 4 episodes
313 (Figs. 2.26, 2.27, 2.28, 2.29, and 2.30).
314 14. Updating the game reward with a badge system and the UI to foster cognitive
315 interactivity: The three-star system did not provide enough information to the
316 payers on their game-based mathematics problem-solving performance. To
Fig. 2.27 A cohesive game world: Desert
Fig. 2.28 A cohesive game world: School
promote reflective inquiry, we used a multi-badge system to provide more 317

detailed feedback on different aspects of the game-based problem-solving per- 318
formance (e.g., fluency, accuracy, and thriftiness). We also refined the user 319
interface again to increase cognitive interactivity. For example, in the collection 320
action, related object information was presented as a tooltip in tables or charts. 321
The log-in was moved to the server to allow online play on Chromebooks 322
(Figs. 2.31, 2.32, and 2.33). 323
Fig. 2.29 A new episode: Farm
Fig. 2.30 Being a 4-episode, 19-level game

Fig. 2.31 A badge system
Fig. 2.32 Support for cognitive interactivity in the collection action

Fig. 2.33 Online login
Fig. 2.34 New training levels
324 15. New training levels and the movement interface: To better reduce the entry
325 learning curve and the technical functionality of object maneuvering in a 3D
326 space, we added more training levels and designed a new item manipulation
327 system. The new system provides unobstructed views on the item’s three-
328 dimensional movement to allow easier manipulation and a higher level of
329 responsiveness. At this point, the game had 43 game levels distributed across 4
330 game episodes (Figs. 2.34, 2.35, and 2.36).
Fig. 2.35 New movement interface
Fig. 2.36 Being a 4-episode, 43-level game
In comparison with the iterative revisions in the gameplay and level development, 331
the game logs have remained relatively stable after the core game actions are speci- 332
fied. This allows for comparison between the older logged play data and the new 333
ones, as well as the refining and validation of the statistical model for stealth assess- 334
ment. Where the game has gotten more complex, the logs have become s impler dur- 335
ing the development process to allow for quicker parsing of data and smaller 336
file sizes. 337
Fig. 2.37 Heat map analysis example
338 Originally, nearly everything was logged. The player’s position was recorded
339 approximately once per second to allow for a heat map analysis of their paths, as
340 you can see in Fig. 2.37. This required an xml block similar to:
341 <Auto Position>

342 <runningTime>2.009376</runningTime>
343 <Position>(-29.0, 0.8, 42.1)</Position>
344 </Auto Position>
345 The log files were flooded with the above entries, which caused problems with
346 file size and parsing, as well as reading them in a glance. In the next revision, we
347 removed the auto position recordings. Still, nearly every action was recorded. The
348 log files (see Fig. 2.38) were more manageable but still required parsing to get to the
349 variables we used in the Bayesian network. We then further trimmed the log struc-
350 ture to only include what we would use in the statistical model. The logs (see
351 Figs. 2.39 and 2.40) then varied by level but could be interpreted in a glance.
352 2.4 Researcher Positionality and Reflection Notes
353 As Van Manen (1990) pointed out, our preunderstandings predisposed us to inter-
354 pret the nature of the phenomenon in a phenomenological inquiry. As part of the
355 effort to bracket these predispositions, we presented a brief introduction of our roles
356 in the design team as well as a reflective description of individual perspectives and
357 perceived experience of the interdisciplinary design process for game-based learn-
358 ing platform development. Our motives, domain-specific backgrounds, and habits
359 of and dispositions toward design certainly impact our reflective analysis of the
360 design process and the findings that emerged as they are part of the lens brought to
361 the project. However, it is important to understand that these preunderstandings did
362 not remain unbracketed during design and research. Additionally, the authors
363 applied peer debriefing and member checks in the inspection and rectification of the
364 data and interpretations.
2.4 Researcher Positionality and Reflection Notes 41
Fig. 2.38 Refined game

log example 1

log example 2

log example 3
365 2.4.1 Fengfeng Ke
366 My belief is that mathematics is a way of thinking, and learning is to develop powerful ways
367 of thinking.
368 Role and responsibility in the design team I am the principal designer of the
369 E-Rebuild project and responsible for leading all facets of the design and research
370 work, including the design, in-field testing, and refinement of E-Rebuild, the man-
371 agement of the project team’s communication and collaboration, as well as the data
372 collection and analysis. As an associate professor of educational psychology and
373 learning system designer at the Florida State University, my design and research
374 interests lie in digital game-based learning and innovative learning environment
375 design. Besides E-Rebuild, I have participated in multiple other research projects on
376 game- and simulation-based interactive learning systems.
377 Personal goal or interest for the E-Rebuild design As a former gamer, I prefer
378 games that involve puzzle solving and construction. As a teacher and a lifelong
379 learner, I believe learning is ultimately self-regulated, and teaching is a process of
380 instilling thinking or lifelong learning skills. These dispositions are naturally infused
381 into my vision for the E-Rebuild design. As a learning system designer, I think the
382 vital role of technologies for learning is not offering an alternative delivery media or
383 a motivation tool but facilitating a new way of understanding and organizing
384 learning. My goal for designing E-Rebuild is hence not creating a single digital
385 learning tool but using it as a test-bed to examine and refine our conjectures of
386 mathematics and learning, to experiment with previous hypotheses governing the
387 integration between learning and play, and to discern new tools and methods to
388 designing an active, meaningful learning experience.
389 Perceptions of interdisciplinary design of E-Rebuild Designing a learning sys-

390 tem is naturally an interdisciplinary process. During previously designed projects, I
391 have worked with content experts, educators, developers, and evaluators from differ-
392 ent fields. These works, however, were more multidisciplinary than interdisciplinary,
393 in that the design and development efforts of team members are frequently segmented
394 by time (e.g., the evaluation won’t start until the functional prototype is ready) or by
395 a clear-cut responsibility division. In spite of the occasional brainstorming or peer
396 debriefing opportunities, the overlapping or truly collaborative design efforts in those
397 projects are usually limited. E-Rebuild project, differently, is an extensive and longi-
398 tudinal project that requires and allows for a full understanding of the discipline-
399 specific subcultures, a gradual development of project-specific design culture, and
400 iterative testing of interdisciplinary perspectives and methods in the design artifact.
401 In spite of the numerous discussions of generic principles of interdisciplinary col-
402 laboration or professional development, I have found the interdisciplinary design of
403 E-Rebuild, by itself, is a learning by doing process. In our earlier project meetings, I
404 felt less comfortable or occasionally lost with the team communications; I wondered
405 whether and how those lengthy interdisciplinary communications would be a worthy
investment to the quality of the final artifact, given a tight project schedule. All these 406
uncertainties, however, got addressed once the mutual understanding of the language, 407
stances, and concepts were achieved and when we gradually found our protocol to 408
channeling differing interests and expertise into the tangible working plan. 409
Perspectives of interdisciplinary design and research for educational game 410

development In spite of a plethora of research on game studies, designing an 411
educational game, like art, is an utmost ill-structured problem. Intuitive design, 412
modifying a worked example, trial and error (or design experimentation), and work- 413
ing backwards (e.g., an analytical or evaluative review of exemplary games to 414
inform the current game design) are frequent strategies adopted by educational 415
game designers. Designers, educators, or evaluators may have an overarching goal 416
or vision about the desirable game-based learning experience, but it is difficult to 417
operationalize such an experience as tangible objectives. Even when objectives are 418
set, the problem space for achieving these game design objectives is murky. Through 419
conducting and reporting a reflective and phenomenological analysis of the core 420
facets and salient patterns emerged from a longitudinal design project, I hope our 421
work will help to illustrate and better define the problem space of interdisciplinary 422
education game design. 423
2.4.2 Valerie Shute 424
Measure twice and cut once! 425
Role and responsibility in the design team My role in the E-Rebuild project is 426
primarily related to all assessment-related aspects of the project. For a brief back- 427
ground, I am the Mack and Effie Campbell Tyner Endowed Professor of Education 428
in the Department of Educational Psychology and Learning Systems at Florida 429
State University. Before coming to FSU in 2007, I was a principal research scientist 430
at Educational Testing Service where I was involved with basic and applied research 431
projects related to assessment, cognitive diagnosis, and learning from advanced 432
instructional systems. My general research interests hover around the design, devel- 433
opment, and evaluation of advanced systems to support learning—particularly 434
related to the twenty-first-century competencies and STEM content. My current 435
research involves using games with stealth assessment to support learning—of cog- 436
nitive and noncognitive attributes. As the originator of the term and technologies 437
surrounding “stealth assessment,” I’m pleased to see it become more broadly 438
accepted and used in new research projects (including this current one). 439
Personal goal or interest for the E-Rebuild design When we began thinking and 440
talking about this project, prior to funding, our goals were to design a platform that 441
integrated architectural design and building with evolving and deepening mathemat- 442
ical understanding. Moreover, this was to be designed and developed in the context 443
444 of a game with which kids would want to engage and learn. For the past 4 years, and
445 across a lot of iterative pilot testing, our original goals are successfully being met.
446 With regard to the assessment part of the story, we have focused on refinements to
447 both the local and server-side game logs as a way to capture relevant gameplay data
448 to inform the math competency model (developed at the onset of the study).
449 Concurrently, we continue to explore the viability of other data mining approaches
450 to supplement the stealth assessment within E-Rebuild. Together, the more t op-down
451 (theoretically driven) stealth assessment approach coupled with the more bottom-up
452 (empirically driven) data mining approach should provide more accurate estimates
453 of competency states than either alone. And having more accurate estimates—avail-
454 able at any time and at various grain sizes—will provide the basis for more targeted
455 and effective math learning supports in the game, including the automatic selection
456 of the next best task to provide to the learner based on his/her current needs.
457 Perceptions of interdisciplinary design of E-Rebuild Our weekly meetings of

458 the entire interdisciplinary team have been, overall, enjoyable and interesting.
459 Because everyone has their own areas of expertise, it is often a very educational
460 experience as well. Furthermore, when a topic is discussed, multiple perspectives
461 are available to look at an issue from all sides. Communication is respectful, even
462 when there are differing views on solving a particular problem.
463 We also have been able to integrate “lessons learned” from other currently funded
464 and directly related research. For example, in one of our current National Science
465 Foundation (NSF) cyber-learning grants using stealth assessment within a game
466 called Physics Playground, we needed to come up with some difficulty indices in
467 order to rate the various levels in the game relative to their difficulty. Originally, we
468 had derived indices related to particular aspects of the level (e.g., if there were a lot
469 of obstacles in the problem that made it harder, and so on). However, we soon real-
470 ized that additionally considering the nature of the physics content related to the
471 game level was necessary to make decisions about “level difficulty” overall. The
472 information gleaned from the cyber-learning project has thus been used in the cur-
473 rent E-Rebuild project. That is, we have developed an approach in E-Rebuild to
474 categorize all of the various game tasks based on a difficulty index that consists of
475 both (1) the complexity of the problem representation and solution processes in a
476 given level and (2) the mathematical competencies required in the solution, as well
477 as prerequisite relationships between the competencies and sub-competencies. This
478 in turn has engendered the development of an adaptive game-based algorithm,
479 which can select appropriate learning tasks based on the real-time estimate of a
480 learner’s competency profile to support the development of focal math knowledge
481 and skills.
482 Perspectives of interdisciplinary design and research for educational game
483 development In summary, working within an interdisciplinary design team has
484 been a great experience. As mentioned, because of the varying areas of expertise, I
485 never leave a meeting without learning something new (and often multiple new
486 things). The discussions of issues among the team members have resulted in an
expansion (and occasionally a refinement) of different areas given fresh eyes on a 487
problem. What has been most enlightening to me is the direct relationship between 488
architectural design processes and math skills, at a general level. So in a sense, suc- 489
cess in the game provides a two-for-one—gaining architectural designing and build- 490
ing skills and associated math knowledge and skills. Another eye-opener is what 491
takes place when we take our game into the schools. Invariably, the teachers and 492
importantly the students have a lot to share with regard to the functionality of the 493
game. Both teachers and students are very clear about what they want and need to 494
succeed—as designers of new levels (teachers using the platform) or as players of 495
the game (students). 496
2.4.3 Kathleen M. Clark 497
A view from a non-gaming, non-assessment, mathematics teacher educator. 498
Role and responsibility in the design team I have just completed my 31st year in 499
mathematics education. I began teaching mathematics in 1987 in an inner-city 500
Miami high school. I taught mathematics in a variety of public schools in Florida 501
and Mississippi until 2001, when I was awarded an Einstein Distinguished Educator 502
Fellowship. During the 1-year fellowship, I worked in educational policy for the US 503
Senate, and though I did not return to my secondary teaching career afterward, I 504
decided to pursue my doctorate in Mathematics Education at the University of 505
Maryland. I have been at Florida State University (FSU) since August 2006. Early 506
in my FSU career, my research focused primarily on pre-service and in-service 507
mathematics teachers, with an emphasis on the role of history of mathematics in 508
mathematics teaching and learning. I currently spend the majority of my time con- 509
ducting research on the use of primary historical sources in the teaching of under- 510
graduate mathematics. 511
My role on the E-Rebuild team is to serve as the “mathematics education expert.” 512
In this role, I focus on providing support for the mathematical content both within 513
the game itself and within the various forms of assessments used in the project. The 514
content that informs the game and the assessments were—in the first development 515
of E-Rebuild—based upon the Common Core State Standards for Mathematics 516
(CCSS-M). However, as the game has developed, we included attention to the 517
Florida Mathematics Standards (FMS), and we do so because the game is currently 518
and primarily tested with Florida mathematics teachers and students, and we must 519
be cognizant of the minute differences between CCSS-M and FMS. 520
Personal goal or interest for the E-Rebuild design I was initially drawn to the 521
E-Rebuild project—particularly from the students’ perspective—because I often observe 522
limited appropriate use of multimedia tools (e.g., appropriate web tools, electronic or 523
Internet-based games, computer software, etc.) to promote middle grade students’ learn- 524
ing of mathematics. I was excited by the potential of E-Rebuild as a means by which 525
526 students would learn or, possibly practice, not only core mathematical concepts (e.g.,
527 ratio and proportion skills, properties of geometric shapes and determining their area,
528 perimeter) but solve problems in a contextually rich and meaningfully engaging game
529 environment.
530 My experience with the E-Rebuild game itself has been one of healthy skepti-
531 cism. I have never been a gamer, so this may contribute to my skepticism, at least to
532 a small degree. That said, I find myself thinking about the numerous variables that
533 have the potential to impact the student learning experience as they play E-Rebuild.
534 For example, are the directions clear for the student in each game level? Is the game
535 visually and structurally appealing to young learners? What are the connections that
536 students make with regard to the mathematical concepts that they meet in the game?
537 What does the context and gameplay environment contribute to students’ engage-
538 ment with mathematical ideas? In my observations of student gameplay, I have
539 noted a high level of student frustration from the perspective of game design fea-
540 tures and not from the perspective of mathematical content. Thus, this prompts me
541 to think about how little expertise I have with regard to game-based learning—or
542 educational games in general—because I wonder about the amount of unproductive
543 struggle students experience with game features and how this may detract from the
544 role of productive struggle when solving the mathematical problems imbedded in
545 the game. And, due to my lack of expertise in “stealth assessment,” I do not under-
546 stand how the components of these more affective experiences with E-Rebuild may
547 factor in to whether students are learning as a result of their E-Rebuild play.
548 Perceptions of interdisciplinary design of E-Rebuild I have participated in a
549 number of interdisciplinary projects. I continue to be energized by the way in which
550 experts and practitioners from different disciplines and fields come together to cre-
551 ate, design, and test instructional tools (e.g., curriculum, materials, interventions) to
552 promote student learning. There are several meeting structures used within the
553 E-Rebuild project: Some are focused specifically on design of tasks, levels, and
554 episodes that comprise the game itself. Other meetings are focused on the inherent
555 mathematical concepts, skills, and problem-solving situations at the core of the
556 game, as well as the components of the architectural design elements that contribute
557 to the underlying mathematics. Regardless of the purpose or goal of any given
558 design meeting, I have found the discussions to be interesting, and they challenge
559 me to try to find ways in which I can learn as much as possible while also drawing
560 on my mathematics education background in order to contribute in optimal ways.
561 I believe the most challenging aspect of the design meetings has been navigating
562 the various communication styles within the group. We often assume we have
563 “taken-as-shared” knowledge, and since we do not, it often takes us several itera-
564 tions to operate with the same (or at least similar enough) knowledge. I have always
565 taken my own notes during the numerous meetings in which I have participated over
566 the years (though I often felt I was the only one—perhaps because I was the only
567 one who needed to!), and these have been helpful in being able to feel that I can
568 keep up with the various aspects of the E-Rebuild work. I think that consistent use
569 of agendas and assigned tasks for regular meetings would have been beneficial. For
example, by scheduling and planning specific tasks and agenda items, I believe a 570
more equitably distributed range of ideas and foci would be addressed in the 571
meetings. 572
Perspectives of interdisciplinary design and research for educational game devel- 573
opment I am intrigued by the notion of what we can learn from the d evelopment and 574
the use of educational games in the learning of mathematics. I think there is much more 575
that we can know (and, consequently, contribute to the field) with respect to student 576
learning and E-Rebuild gameplay. However, as a qualitative researcher, I believe that 577
we must continue to conduct student and teacher interviews to better understand the 578
student experience with E-Rebuild. Additionally, the various teacher experiences are an 579
important part of this inquiry. Students have played E-Rebuild as part of the regular 580
mathematics classroom activity, in non-mathematics classrooms, and within after 581
school and summer camp contexts. I believe it is wise to investigate the ways in which 582
students have played E-Rebuild in these different formal and informal contexts and 583
work to understand the assessment results with these different implementation lenses 584
in mind. I think there is great potential for the results of these investigations to assist us 585
in focusing on further game design and development to maximize the potential for such 586
a tool to improve middle grade students’ nonroutine problem-solving abilities. 587
2.4.4 Gordon Erlebacher 588
In my view, a perfect educational game should improve player skill (in an idealized situa- 589
tion) without the player being aware of the purpose of the game. Not unlike good advertis- 590
ing which influences the actions of a person without that person being aware of the ad. 591
Role and responsibility in the design team The E-Rebuild project team was 592
formed in 2011, and I participated from the onset. My role in this project is to ensure 593
that the team applies sound game design principles during the development of the 594
game and help manage implementation of its game mechanics. Prior to my arrival 595
at FSU, I was a researcher at NASA Langley and at the Institute for Computer 596
Applications in Science and Engineering. During that period, I conducted theoreti- 597
cal and computational research in the field of turbulence. In 2009, I created a first 598
course on Game Design at FSU, which attracts undergraduate and graduate students 599
over a wide range of science and nonscience departments. I have often felt that 600
properly designed games could help students achieve more of their potential and 601
mathematics is a natural subject matter given my background. I am currently Chair 602
of the Department of Scientific Computing at Florida State University, an interdis- 603
ciplinary department whose mission is to develop computer algorithms with appli- 604
cations in the applied sciences. Since 1996, I have been involved in computational 605
science, developing algorithms in the fields of visualization, GPU programming, 606
flow simulation, and more recently in computational neuroscience and deep learn- 607
ing. Over the past 30 years, I have been exposed to many frameworks, software, 608
computer languages, and technologies, which has helped me develop the skills to 609
interact with researchers in different fields. 610
611 Personal goal or interest for the E-Rebuild design In the early stages of the project,
612 I envisioned that we would eventually develop an exciting and engaging game that
613 middle-school students would want to play and as a result undergo a gradual improve-
614 ment of their mathematical skills. The game would adapt to the player, choosing sce-
615 narios and tasks commensurate with the player’s ability. This would be accomplished
616 through the collection of data (player movement, mouse clicks, imagery) that would
617 then feed into an assessment system that would measure the mathematical proficiency
618 of the player. Together with player history, new scenarios would be generated, adapted
619 to each player. Through a team effort, the game evolved somewhat differently. School
620 teachers would be responsible to creating the scenarios and tasks appropriate for their
621 students, and the players would be aware that the game was meant to improve their
622 mathematical skills. We have been refining the competency model and field-testing the
623 game to validate the model and ensure that the students are indeed improving through
624 gameplay. At this time, my goals are more ambitious. The difficulty of finding enough
625 participants to test the game and model leads to the idea of creating a synthetic player
626 based on machine learning principles with a “brain” that can adapt to gameplay.
627 Although this brain would be a poor facsimile of a human brain, my hope is that it
628 would learn in different ways and at different rates depending on the sequence of tasks
629 and scenes presented, and this would lead to insights into further game improvements.
630 At this stage, this remains just a vision for future research.
631 Perceptions of interdisciplinary design of E-Rebuild Early in the project, after

632 the project was funded, the whole project team met on a weekly basis to discuss the
633 nuts and bolts of the game, such as game mechanics and backdrops, along with
634 some initial tasks the player would have to complete. These meetings were fun but
635 challenging in the early days because of the different backgrounds and past experi-
636 ences of the participants, which ranged from architecture to educational games to
637 instructional systems to mathematics education to assessment and science computer
638 simulations. Our different past experiences led to rather different ideas on what
639 would go on in the mind of an eighth grader, which in turn would impact decisions
640 related to game development. Discipline vocabulary, which differed between disci-
641 plines, also impeded progress on occasion, and it behooved us to allow for differ-
642 ence of opinion which was not necessarily substantive but could be the result of a
643 different understanding of word usage. A recent case in point was confusion in the
644 use of the term “elevation view,” which for an architect refers to a side view but to
645 a game designer might mean view from above or a third-person view. Individual
646 members also had a unique vocabulary, which sometimes/often led to entire discus-
647 sions when different people used the same word in discussions, but with different
648 meanings. Once, the majority of a meeting was spent debating the meaning of the
649 words task, scene, objective, and how to use them to describe different components
650 of the game. Another early discussion revolved around the task of construction;
651 should objects be controlled via the keyboard or using the mouse? How would the
652 player respond to these different choices? How much help should the player be
653 given and in what form should it be made available. Should the help be in-game or
654 out-of-game? Should the game flow be interrupted or should we maintain player
References 49
immersion into the game. There was disagreement on all of these aspects, which 655
eventually were resolved. The cases that remained open were decided via testing by 656
the students. Overall, I enjoyed the meetings, gaining an appreciation of other fields 657
of study and how they related to ideas presented. 658
Perspectives of interdisciplinary design and research for educational game 659

development Interdisciplinary research and interdisciplinary design are concepts 660
that have been around and promoted for many years. From my personal experience, 661
the E-Rebuild project is probably the most satisfying interdisciplinary effort in 662
which I have participated. Very often, participants meet once a year or biannually 663
and share what they have accomplished since the previous meeting. On the other 664
hand, our whole-team meetings were weekly initially and, more recently, biweekly. 665
All ideas are shared and discussed, and progress made since the previous meeting is 666
presented. I have learned a lot from our interactions, not only in terms of game 667
design but also on the personal level. I have learned to appreciate some of the tools 668
used by architects, educators, and programmers to help others learn and enjoyed the 669
process through which different goals eventually merged into a game that pleased 670
the entire team. Eventually, it became clear that stated goals during our meetings 671
often reflected the background of the person. For example, our architect pushed for 672
a game that was much more architecturally oriented in terms of realism, while oth- 673
ers on the team were more interested in gameplay at the expense of realism. Good 674
teamwork produces a game that compromises between these different visions with- 675
out sacrificing the vision of the team leader. For large projects, a truly interdisciplin- 676
ary approach is often indispensable, particularly when a game touches upon many 677
different disciplines as it does in our case (construction, education, mathematics, 678
and simulation). 679
References 680
Almond, R. G., Mislevy, R. J., Steinberg, L., Yan, D., & Williamson, D. (2015). Bayesian networks 681
AU4 in educational assessment. Springer. 682
Breuer, J., & Bente, G. (2010). Why so serious? On the relation of serious games and learning. 683
Journal for Computer Game Culture, 4(1), 7–24. 684
Butler, E., Andersen, E., Smith, A. M., Gulwani, S., & Popović, Z. (2015). Automatic game pro- 685
gression design through analysis of solution features. In April (Ed.), Proceedings of the 33rd 686
annual ACM conference on human factors in computing systems (pp. 2407–2416). ACM. 687
Dewey, J. (1929). The quest of certainty: A study of the relation of knowledge and action. 688
New York: G.P. Putnam’s Sons. 689
Eriksson, I., & Lindberg, V. (2016). Enriching ‘learning activity’ with ‘epistemic practices’– 690
enhancing students’ epistemic agency and authority. Nordic Journal of Studies in Educational 691
Policy, 2016(1), 32432. 692
Giorgi, A. (2009). The descriptive phenomenological method in psychology: A modified Husserlian 693
approach. Duquesne University Press. 694
Groenewald, T. (2004). A phenomenological research design illustrated. International Journal of 695
Qualitative Methods, 3(1), 42–55. 696
697 Hiebert, J., Carpenter, T. P., Fennema, E., Fuson, K., Human, P., Murray, H., et al. (1996).
698 Problem solving as a basis for reform in curriculum and instruction: The case of mathematics.
699 Educational Researcher, 25(4), 12–21.
700 Hunicke, R., LeBlanc, M., & Zubek, R. (2004, July). MDA: A formal approach to game design
701 and game research. In Proceedings of the AAAI Workshop on Challenges in Game AI (vol. 4(1),
702 p. 1722–1726). AU5
703 Mislevy, R. J., & Haertel, G. D. (2006). Implications of evidence-centered design for educational
704 testing. Educational Measurement: Issues and Practice, 25(4), 6–20.
705 Rodriguez, H. (2006). The playful and the serious: An approximation to Huizinga’s Homo Ludens.
706 Game Studies, 6(1), 1604–7982.
707 Van Manen, M. (1990). Beyond assumptions: Shifting the limits of action research. Theory Into
708 Practice, 29(3), 152–157.
709 Weingart, P. (2000). Interdisciplinarity: The paradoxical discourse. In P. Weingart & N. Stehr
710 (Eds.), Practising interdisciplinarity (pp. 25–41). Toronto, ON: University of Toronto Press.
Author Queries

AU2 Please confirm the inserted citation for Figs. 2.1–2.36.
AU3 Please provide better quality artwork for Figs. 2.10, 2.11, and
2.18.
AU4 Please provide publisher location for “Almond et al. (2015),
Giorgi (2009) and Butler et al. (2015)”.
AU5 Please provide publisher name and location for “Hunicke et al.
(2004)”.
Chapter 3 1
Interdisciplinary Design Activities 2
and Patterns 3
Abstract In this chapter, we provide a reflective and analytical description of the 4

interdisciplinary design activities of E-Rebuild, identify driving design questions 5
and salient design patterns that capture and frame the essence of E-Rebuild develop- 6
ment, and discuss distilled meta-generalizations that help to decompose the inter- 7
disciplinary learning game design processes to inform future work related to design, 8
research, and deployment. The description of salient patterns of interdisciplinary 9
game design in this chapter provides a contextualized design narrative/account of 10
core design processes along with an analytical synthesis of core design pattern ele- 11
ments—a design problem statement with its context and specifics, the solution or 12
technique to solving the stated problem, and the pattern of transferring or scaling 13
this design solution or design move. 14
3.1 Introduction 15
AU1 There has been a growing recognition that the process of designing and developing a 16
computer-supported, highly interactive learning system in general and games or simu- 17
lations for learning in particular, is by its very nature an interdisciplinary and challeng- 18
ing task. Prior works of design science and serious game design highlight the importance 19
for learning and game designers to record, identify, and distribute not only detailed, 20
contextualized design narratives but also higher-level, distilled design knowledge such 21
as design patterns (Pratt, Winters, Cerulli, & Leemkuil, 2009; Winters & Mor, 2008). 22
A design pattern is a high-level specification for a solution methodology to a 23
design problem, by specifying the particulars of the problem and highlighting recur- 24
ring solutions that are field tested in real-world application development. In this 25
chapter, we provide a reflective and analytical description of the interdisciplinary 26
design activities of E-Rebuild, identify driving design questions and salient design 27
patterns that capture and frame the essence of E-Rebuild development, and discuss 28
distilled meta-generalizations that help to decompose the interdisciplinary learning 29
game design processes to inform future work related to design, research, and 30
deployment. 31

https://doi.org/10.1007/978-3-030-04339-1_3
52 3 Interdisciplinary Design Activities and Patterns
32 The design meeting notes, sketches, aids, paper and electronic prototypes, as
33 well as conversation emails during the E-Rebuild project have all been archived.
34 Iterative and reflective interviews were also conducted among design team mem-
35 bers. Adopting the phenomenological qualitative research approach (Moustakas,
36 1994), we have conducted a thematic analysis to explore salient themes or patterns
37 that capture the nature and features of design choices and actions by the interdisci-
38 plinary design team. While coding salient themes emerging from the data, we have
39 also referred to the prior research in the field of design science when seeking the
40 boundary and meaning of a theme or pattern. The description of salient patterns of
41 interdisciplinary game design in this chapter provides a contextualized design nar-
42 rative/account of core design processes along with an analytical synthesis of core
43 design pattern elements—a design problem statement with its context and specif-
44 ics, the solution or technique (with its key structure or mechanism) to solving the
45 stated problem, and the pattern of transferring or scaling this design solution or
46 design move.
47 During the design of E-Rebuild, a variety of driving questions or problems initi-
48 ate the negotiation of the design plans and the coordination of the design knowledge
49 among interdisciplinary team members. It then takes interdisciplinary cases/exam-
50 ples exploration, infield testing, and iterative refinement for the team to settle on a
51 design solution. Moreover, design leadership during the decision-making takes on
52 an important role in the project management, especially within a constrained
53 timeline.
54 3.2 Core Design Patterns and Activities
55 3.2.1 D
efining the Design Goal: A Trinity of Architectural
56 Simulation, Mathematical Learning, and Gaming
57 Problem statement During the initial design meetings of E-Rebuild, interdisci-

58 plinary team members debated and went back and forth to better understand
59 E-Rebuild as “an architectural simulation game for math learning.” Some notable
60 selections from the debate include statements such as: “Should it be mainly an
61 architectural building simulation that provides an architectural design education
62 while embedding mathematical conceptual development at the same time?
63 Should it focus on the representation of mathematical concepts and problems,
64 with architectural scenarios being a meaningful context for mathematical repre-
65 sentation? Or should it prioritize the gameplay and fun elements situated in the
66 practice of architectural and mathematical puzzle solving?” These debating top-
67 ics, in general, revolve around an innate design problem for learning game cre-
68 ation, What is the creative stimulus or the ultimate inspiration for the creation of
69 E-Rebuild as a trinity of content learning, simulation, and gaming?
3.2 Core Design Patterns and Activities 53
Exploration of the design solution The aforementioned design problem roots in 70

the typical challenge of designing a learning game—to integrate learning into core 71
game elements while not violating or corrupting what is enjoyable about games 72
(Garris, Ahlers, & Driskell, 2002; Ke, 2016). Prior research argued that the prospect 73
of using a game to motivate learning lies in the assumption that this game will 74
involve learners in what is fundamentally engaging about the subject, via either a 75
dynamic representation or the simulation of the most “fun” (dynamic) component 76
of the academic domain (Ke, 2016). Yet it remains debatable as to whether and how 77
pursuing “fun” or gameplay as the preliminary design inspiration will drive the uni- 78
fied representation and simulation of content and content-relevant epistemic prac- 79
tice. For example, the following design notes documented a design goal motivated 80
primarily by the gameplay exploration: 81
We will start by building a “Block” level in which the player will select, drag, and drop 82
the “block” object to build a specific architecture form. The process is somewhat akin to a 83
3D version of block stacking or jigsaw game – the player must “fill up” an empty, 3D 84
architect using the building blocks (i.e., revive a building). Task assessment proceeds by 85
testing whether and how much the destined 3D architectural space has been occupied, 86
whether this constructed 3D architecture will collapse with a shaking landscape, etc. 87
The embedded math concept can be geometric shapes, area/perimeter, and 88
transformational geometry (rotation, reflection, and dilation), which can be integrated 89
with the organizational principles of architectural design (such as symmetry or balance, 90
hierarchy, and rhythm). 91
The notes portray a design inspiration that seeks a gameplay that can unify or align 92
with mathematical content and math-related architectural practices. Such a gaming- 93
driven design exploration, however, has been found by the content experts and 94
instructional designers in the team as incapable of capturing the depth and multifac- 95
eted nature of content learning. It prioritizes certain mathematical competencies or 96
levels of learning (e.g., composing and decomposing geometric forms for concep- 97
tual understanding) while ignoring the others (e.g., problem-solving that involves 98
area, volume, or unit rate computation). 99
As an alternative and refinement, the team shifted toward domain content and 100
architectural design as the dynamic starting points. In particular, we started by 101
developing graphical competency models to outline the structure or typology of 102
targeted mathematical competencies. Each node in the math competency model rep- 103
resents a performance objective and highlights learning actions. Exemplary mathe- 104
matical context problems for each node were also gathered. The practice of 105
composing child nodes or decomposing a parent node along with their supporting 106
mathematical problems has helped us conceptualize varied tasks or quests with cor- 107
AU2 responding actions within the game (Fig. 3.1). 108
AU3
Meanwhile, by referring to a pre-developed Architecture and Children Pilot 109
Curriculum by Dr. Anne Taylor (our architecture and design education expert in 110
the team), we drafted an Architecture Terms and Principles document that gathers 111
basic architectural concepts and skills that may relate to mathematical thinking. 112
Fig. 3.1 An example of graphical, competency models
113 This 24-page working document, illustrated in Figs. 3.2 and 3.3, provides
114 definitions and descriptions with rich visuals and real-life examples. Similar to
115 the mathematical competency model, this manual also highlights the componen-
116 tial performances and actions when synthesizing the skills architects and design-
117 ers depend on. Given these features, this manual has worked as a handy manual
118 to frame not only the task development but the game world and object design in
119 E-Rebuild.
120 The simultaneous exploration of the targeted subject domain and content-relevant
121 epistemic practices has helped the project team better delineate the design problem
122 space—an externalized representation of the multifaceted design goal of E-Rebuild.
123 Given this clearly defined, action- or performance-themed design problem space,
124 the team then began anew and were productive in searching, selecting, and assem-
125 bling the necessary gameplay components (e.g., core actions, rules, and backdrop
126 missions) to further frame the design goal and the problem space from a “gaming”
127 perspective.
128 Design pattern summary Defining the design goal of the simulation-based learn-
129 ing game is the delineation of a multifaceted, interdisciplinary design problem
130 space. The framing of the problem space is initiated by a synchronized modeling of
131 domain-specific competencies and the cataloging of simulated epistemic practices.
132 Modeling and cataloging should highlight the fundamental performance expecta-
133 tions and supporting actions for content representation and practice simulation. The
134 competency-driven, action-themed design problem space will then drive the explo-
135 ration and selection of gameplay that captures and unifies content learning and
136 simulated epistemic practices.
TABLE OF CONTENTS
1. THE ORGANIZING PRINCIPLES OF DESIGN
2. GEOMETRIC FORMS IN ARCHITECTURE
3. ARCHITECTURAL VOCABULARY – STRUCTURES
4. SPATIAL RELATIONSHIPS
5. DESIGN AND VISUAL VOCABULARY

(a repetitive process for design)
6. LANDSCAPE DESIGN
7. SKILLS ARCHITECTS AND DESIGNERS NEED AND USE
2. GEOMETRIC FORMS IN ARCHITECTURE

Point (Zero D) - a precise location or place on a plane; usually represented by a dot.
Line (1D) - a geometrical object that is straight, infinitely long and infinitely thin.
Plane (2D) - a flat surface that is infinitely large and with zero thickness.
Volume (3D) - volume is the measure of the amount of space inside of a solid figure.
Circle - a2 -dimensional shape made by drawing a curve that is always the same distance from a
center.
Sphere - a3 -dimensional surface, all points of which are equidistant from a fixed point.
Figs. 3.2 and 3.3 Table of content and example pages of the “Architectural Terms and Principles”
document
Figs. 3.2 and 3.3 (continued)

3.2.2 S
etting Core Gameplay: Meaningful Mechanics that 137
Connect to Intellectual Learning and Artifact Design 138
Problem statement The subsequent design question that receives considerable 139
discussion takes the form, “What will a player do in the game, and why is that 140
engaging and intelligent?” In other terms, what and how will the core game 141
mechanics of E-Rebuild connect intellectual content and architectural design to 142
motivate and guide math conceptualization and problem-solving? Prior research 143
suggests that learning occurs only when play experience connects to intellectual 144
content; yet the process that leads the attainment of an integral and continuing 145
relationship between gameplay and the content to be learned remains murky 146
(Habgood & Ainsworth, 2011). Moreover, there had been some divergence within 147
the E-Rebuild team as to the share of (architectural) design knowledge relates to 148
mathematical one to be embodied and activated by the gameplay actions and rules. 149
Exploration of the design solution We have reviewed exemplary design cases in 150
relevant disciplines and explored discipline-specific inquiries or tasks when select- 151
ing the core game mechanics of E-Rebuild. The discipline-specific design cases and 152
ideas were typically presented by a discipline expert, reviewed and commented on 153
by the team at design meetings. Shared and prominent ideas emerging from these 154
reviews and discussions were then aggregated to be paper prototyped as the initial 155
version of gameplay. 156
Treating architectural design as the epistemic practice of mathematical concep- 157

tualization and problem-solving, the team started by listing the key elements and 158
actions of architectural design and those of mathematical reasoning. The team also 159
searched for architectural design and construction cases that inspire the selection of 160
salient actions for intellectual gameplay, as illustrated by the following design notes 161
documented after several rounds of architectural design case discussion. 162
Can we simulate the “building block” and landscape-management design ideas 163
from the New Zealand earthquake rebuild project (https://www.psfk.com/2012/12/ 164
shipping-container-mall.html)? 165
• Building blocks (tools: math tools, design tools, building tools, shape/texture). 166
• Building actions: Navigating/exploring, drop/drag/customize, watch/predict/ 167
reflect, drawing and modeling (outside the game or using mixed reality)? 168
• Building tasks (reflecting a bottom-up approach, integrating Anne’s ideas and 169
content in the Architecture and Children Pilot Curriculum): 170
Architectural floor plan design—Bubble diagram and a related task: Designing 171
a fast food vegetarian restaurant (spaces only, without equipment) with pre- 172
design exercise (such as decorating a shoe). 173
Structure in architecture (arch, triangle, asymmetrical vs balanced) and a related 174
task: Building a space frame structure for the roof of the picnic shelter with 175
pre-design exercise (e.g., toothpick puzzles). 176
177 Form in architecture or architecture in nature (may involve measurement, data

178 interpretation, scale, and proportion) and related tasks: Using organic struc-
179 tures in nature in a “fantasy” architectural structure for the “moon,” with
180 pre-design puzzle (scavenger hunt for basic geometric shapes and forms in
181 nature) or designing a “people’s” bridge across the canyon to build a jogging
182 and bike path (designing and modeling natural, beam, suspension, or arch
183 bridges).
184 Visual design, creative design.
185 Landscape, city planning.
186 • Building tasks (reflecting a top-down approach).
187 Building blocks and site planning—Using given building objects (e.g., shipping
188 containers) to complete the construction of a site (e.g., a village). This exer-
189 cise involves the organizational principles of design (e.g., Symmetry, Balance,
190 and Proportion) and a basic understanding of the geometric concepts of
191 shape and transformation (e.g., reflection, rotation, and dilation).
192 Reference: Architectural Ordering Principles (document is uploaded to the
193 project site)
t2.1 Architectural
t2.2 system Math system Gameplay element
t2.3 Primary elements: Geometry: Reviving an architecture demolished during
t2.4 Point, line, plane, 1D to 2D to 3D earthquake—Jigsaw game: Recomposing an
t2.5 volume shapes architectural structure via the scattered elements/pieces
t2.6 Area, perimeter,
t2.7 and volume
t2.8 Properties of forms: Geometry: Scavenger hunt game: Collecting building items in
t2.9 Shape, size, color, Shapes varied forms and properties
t2.10 texture Area, perimeter,
t2.11 Position, and volume
t2.12 orientation, visual Proportion
t2.13
194 inertia
195 As the above design note portrays, a real-life post-quake rebuild story, including
196 architectural building elements, actions, and tasks, have driven and inspired the
197 selection of applicable types of gameplay. In addition, a purposeful exploration of
198 the association between architectural and mathematical elements leads to the selec-
199 tion of gameplay that helps unify the math learning and architectural building
200 actions.
201 On the other hand, the learning system designers and game programmers in the
202 team were concerned that the performance of certain architectural building ele-
203 ments, such as creative and visual design or landscape and city planning, were not
204 only beyond the scope of the design goal but also hard for the digital gaming system
205 to gauge. There was also a concern that a variety of differing game mechanics
206 derived from the simulation and representation of varied architectural building tasks
and principles would lead to a lack of focus and consistency in gameplay and hence 207
diminish engagement. Instead, members of the team proposed that only one or two 208
gaming actions be deployed, as suggested in the following design note: 209
A player is given a set of materials and must create a structure. Performance is evaluated 210
in the game by the number of waste materials left over, time to completion, and structural 211
soundness of the structure. Peers can also rate the aesthetic beauty/originality of the 212
structure. There can be several constraints in E-Rebuild, such as 1) amount of materials, 213
2) strength of materials, 3) weight of materials, 4) variety of materials, 5) time, or 6) 214
building specifications. Let’s take constraints 1 and 4. In this case, assume that the player 215
is given 1000 cubic feet of oak wood and 100 square feet of glass. Based on these 216
constraints, the player must decide how to build a house. 217
This game design corresponds to a measurement core mechanic. A person must 218
create a blueprint of the structure that will be used to cut the materials to be used for the 219
structure. This blueprint specifies the dimensions of every element in the structure. A 220
player sketches a blueprint with a pencil and paper outside the game. With the blueprint 221
created, the player enters every unique component (e.g., material, quantity, dimensions) of 222
the structure into a menu within the game. An exemplary entry might take the form: wood 223
board, 50, 2 X 4. The material is always determined by the materials given in the problem 224
(wood, glass). Then the game automatically “cuts” the materials. I do not think the player 225
should cut material as this would get very tedious and repetitive. Once the materials are 226
cut to size, the player must assemble the materials using the mouse and the blueprint. 227
In this design proposal, the core game mechanic is measurement, which consists of 228
the secondary mechanics of structure planning/sketching, material cutting, and 229
material composing. The core gameplay rules are design constraints such as effi- 230
ciency, thriftiness, and soundness. These ideas support simplicity and intrinsic con- 231
tent integration. Yet the practice of a major gaming action (e.g., blueprint sketching) 232
outside the game world may lead to the interruption of game flow. The menu input 233
for the cutting action led to a “design” action that felt like a mathematical drill that 234
was tedious. 235
By integrating the aforementioned two interdisciplinary perspectives, we decided 236
to focus on three core architectural design actions and constraints that identify and 237
unify the most dynamic processes of architectural building and mathematical rea- 238
soning and are in-game play. These core actions included (a) gathering materials 239
(including site surveying, material collection, and transformation), (b) crafting 240
(measurement + cutting), and (c) building (with given design constraints, evaluated 241
according to thriftiness and sturdiness). These core actions and rules defined the 242
first version of E-Rebuild, which was prototyped, user tested, and evaluated based 243
on the feasibility, playability, and learning-necessitating potential. After iterative 244
prototyping and user testing, the material crafting action was later replaced by a 245
material-trading action; an allocation (of space) action was added later. 246
Design pattern summary Gameplay exploration consists of identifying gaming 247

actions and rules that (a) discern and unify the most dynamic and engaging pro- 248
cesses of the performance systems to be simulated, (b) are feasible (e.g., being actu- 249
ated and assessed in the game), and (c) are streamlined and scalable (to compose 250
diverse tasks or scenarios). The gameplay exploration is driven by sharing, review- 251
ing, and aggregating interdisciplinary task examples and design cases. A purposeful 252
253 mapping of the association between applicable game mechanics and the design
254 cases/tasks proposed, along with the subsequent prototyping and user testing, will
255 then help to classify and consolidate the game mechanics.
256 3.2.3 D
esign Constraints for Rebuilding: Evaluating
257 and Rewarding Performance in the Game
258 Problem statement A design inquiry related to the exploration of gameplay

259 defines the evaluation of gaming success. Some team members argued that players
260 require flexibility with their design tools and goals, while others pointed out that
261 specifying design constraints and measurable rules is critical for computerized
262 assessment of a design artifact in the game. A variety of relevant questions arose
263 during the team discussion: What are the success criteria for post-quake rebuilding?
264 Restore (replicate) the pre-quake structures/models or withstand earthquake in dif-
265 ferent magnitudes? Should the rebuilding task be accelerated given a forthcoming
266 earthquake, or self-paced with the option of testing the structure with an earthquake
267 simulation? Do we want to do computerized or peer evaluation? What will be the
268 design constraints and performance evaluation criteria?
269 Exploration of the design solution For the inquiry of rebuilding performance
270 evaluation, the architecture and scientific computing experts led the team in a
271 document research and case review of the earthquake architecture. This exercise
272 contributed a list of factors that mediate the challenging nature of a building task
273 (in terms of sturdiness of the structure against the earthquake vibration), includ-
274 ing the types of earthquake (and hence the necessity of designing diverse support
275 structures) and other associated design parameters (e.g., liquidation issue of the
276 terrain). After the initial prototyping and user-testing of the earthquake architec-
277 ture, we found that simulating all these factors was not only time-consuming but
278 also led to learning challenges far beyond the targeted competencies. For exam-
279 ple, creating an anti-earthquake building would require a player to define the
280 center of mass of a composite structure and calculate friction. Evaluating the
281 collapsed objects to gauge the structure’s sturdiness is also tricky. For example, it
282 is difficult to define the degree of “collapsing.” What would be the consequence
283 if the structure only has one layer and hence would only rotate or move instead of
284 falling down given the earthquake vibration? Given all these design consider-
285 ations, the team decided to focus on structure/model rebuilding rather than build-
286 ing anti-quake structures.
287 In the initial design phase, the team could not persuade each other and hence
288 came to a compromised decision that we would enable both approaches of evaluat-
289 ing a gamer’s design performance: (1) computer evaluation based on the structure
290 or functionality of the architecture (e.g., whether the architecture will collapse with
291 shaking, how much the site space is filled and used efficiently) and (2) evaluation by
peers on the organizing principles and/or art/beauty of the architecture (e.g., peer 292
critique, or design-protocol-based, portfolio-based evaluation), which can promote 293
reflection on action. Yet with a clearer definition of the design goal in the later phase 294
and after more elaborated discussion on the logistics and training required for 295
implementing peer evaluation with middle-school students, the team collectively 296
decided to drop the peer evaluation idea. 297
By synthesizing the reviewed cases and documents, the architect and learning 298
environment designers in the team co-proposed a list of the design parameters act- 299
ing as the performance evaluation standards for E-Rebuild gameplay. These param- 300
eters then drove the design of the game reward mechanism, as the following figure 301
illustrates (Fig. 3.4). 302
Design pattern summary The gameplay rules, which involve mainly the evalua- 303
tion of game performance and the game reward mechanism, are confined by the 304
success criteria of the simulated epistemic practice and the predefined design prob- 305
lem space. Research of the exemplary cases in the subject domain will contribute 306
evaluation rules that are aligned with the authentic performance success, while the 307
review of the predefined design goal and prototyping along with infield user testing 308
will help to delimit goal-oriented and feasible rules. 309
3.2.4 D
esign Execution: Grounding Ideas in the Computer 310
Game System 311
Problem statement The execution or implementation of design propositions via a 312

computer game engine involves grounding and concretizing sketchy ideas related to 313
the user interface, tools, functions, interactive objects, landscape, logging structure, 314
narratives, and scaffolds/aids. The processes and end products that result from 315
developing these game system components are interdisciplinary in nature, convey- 316
ing and testing varied interpretations by the team members of game-based and 317
Fig. 3.4 Design parameter and game reward mechanism

318 architectural design-oriented learning engagement. These salient design execution

319 processes in E-Rebuild evolved around a series of prompting questions or debatable
320 perspectives, such as the following:
321 • How will the player actually move materials around? What is the most intuitive
322 interface for rotating, stacking, and customizing (e.g., cutting and scaling)
323 objects? What exemplary interface mechanisms can act as a reference?
324 • Avoid text entry, if possible.
325 • Will “model sketching?” be handled from within or outside the game environ-
326 ment? Or should we make a set of model sketches (simulating real models from
327 Google Earth via Google SketchUp) available to the player and then perhaps
328 have the player construct model sketches at advanced levels?
329 • How will NPC characters be designed? For example, should we embed a “mini-
330 me” avatar of the player to promote identity development and improve under-
331 standing of the proportion concept?
332 • Is it possible to embed an artificial intelligent agent who can scaffold game play
333 emotionally and cognitively?
334 Given the above questions, the teams’ perceptions varied and could be character-
335 ized as four discipline-related dispositions which prioritized, respectively, (a) intu-
336 itiveness in gaming interactions and hence an increased opportunity of game
337 engagement, (b) motivated practice of mathematical reasoning and computation, (c)
338 increased opportunities for (architectural) design thinking, or (d) sufficiency and
339 quality of evidence collection to drive game-based learning assessment. These
340 diverse perspectives triggered lengthened, controversial discussions on the format
341 of a user interaction widget (e.g., numerical value entry, or multiple choice, versus
342 scrollbar), granularity or fidelity level in the simulation of core design actions (e.g.,
343 building, model sketching) in relation to the prospect of facilitating math or design
344 learning, the design of game world and objects, the proportion and framing of nar-
345 ratives in the game world, the approach of logging and assessing gaming actions,
346 and the approach of embedding in-game support (e.g., background help versus
347 active prompts).
348 Exploration of the design solution The uncertainty in the design execution of
349 E-Rebuild was resolved and settled via iterative design prototyping, expert review,
350 and infield user testing. Feasibility, learnability, and playability are three integral,
351 yet sometimes controversial, facets to be balanced when we gauged and settled on
352 game development details.
353 The interaction interface for core gaming actions (e.g., to build item maneu-
354 vering) was debated in relation to whether activating physics during the building
355 actions would be beneficial or disruptive to gameplay and learning engagement. On
356 the one hand, positioning, stacking, and rotating 3D physical objects are more chal-
357 lenging than maneuvering nonphysical ones. With the physics law activated, objects
358 can bump off each other, fall down, and crash the structure. Thus, building a struc-
359 ture becomes a carefully planned and precisely executed series of actions and
movements. Such an interface has created a learning curve and a reduced sense of 360
autonomy for game players, especially those not used to 3D gaming. On the other 361
hand, simulating the authenticity of the building process is a natural part of an 362
architectural simulation game. We also observed that by decreasing the intuitive- 363
ness of object maneuvering interaction, the game has made it compulsory for a 364
player to engage in planning their building movements (e.g., where to position each 365
item and how to rotate it precisely for stacking) and hence reinforced their spatial 366
and geometrical reasoning. Besides, adding a “snapping together” feature to the 367
stacking function has made layering an object on top of another more efficient. In 368
other terms, the team compromised on a trade-off and purposeful integration 369
between learnability and playability during the gaming interface development. 370
Another example of compromise appeared in the decision to adopt a numerical 371
value entry, rather than the choice button or scrollbar, for user input. The human- 372
computer interaction, and gaming literature in general, suggests that text entry may 373
interrupt the game flow. Yet our initial user testing of the choice button and scrollbar 374
for core gaming actions (such as training items and allocating resources) indicated 375
that they were associated with frequent exertion of guessing and random clicking, 376
and the lack of mindful mathematical computation. Learning system designers and 377
educational measurement experts then argued that prompting players to enter spe- 378
cific numerical values would not only increase the chance of mathematical thinking 379
but also provide more direct evidence for game-based learning assessment. This 380
proposition was executed and infield tested. The team agreed that an intermediary 381
interface (e.g., text entry) helped to necessitate game-based learning engagement. 382
Tool development for the simulated design practice and mathematical problem- 383
solving in E-Rebuild was another salient process experiencing iterative review and 384
refinement. For example, a cutting/scaling tool that “cut/scale items along the three 385
coordinate direction x, y, and/or z” was proposed as the following development note 386
outlined: 387
Scale and cutting tools: Using a drag-and-highlight tool to click on the vector points of 388
the original object (click the right area, then the area would glow to show the tolerance) 389
to define the x, y, z value of the target cube, or to define the radius of the target sphere, or 390
the other relative key values of the other geometry forms, to have it scaled or cut. 391
This tool aims to represent various mathematical concepts dynamically. Specifically, 392
the cutting interaction would assist the conceptual development in composing/ 393
decomposing geometric forms. The scaling interaction would assist proportional 394
reasoning and help correct a misconception that division (multiplication) does not 395
always create a smaller (bigger) element. Yet when prototyped and infield tested, we 396
found that this cutting/scaling tool led to incoherent game play, due to (a) the diffi- 397
culty of spawning irregular shapes/figures (e.g., cutting a pyramid or a polyhedron), 398
(b) its conflict with the trading mechanic that aims to regulate what players can buy 399
(since they could buy a big item and cut/scale it), and (c) the potential of breaking 400
the thrifty rule (because scaling creates the potential to waste material). The team 401
debated as to whether the tool should be removed. It was considered a powerful tool 402
facilitating mathematical reasoning and learning, yet difficult to be fully executed 403
and not aligned with the core game mechanics. The final decision was that the tool 404
405 would be activated only for particular levels/tasks wherein item customization
406 would be aligned with other gaming actions and work as a reward of successful
407 game play to motivate purposeful geometrical reasoning.
408 The “highlighter” was another tool streamlined during the development phase.
409 This tool, originally proposed to enable quick object “positioning” and integration
410 of architectural design/planning, was built based on a concept of grid view by refer-
411 ring to the editing interface of Portal 2 Puzzle Maker. The following development
412 note is a brief outline of object positioning action using a highlighter tool:
413 (The player) define the construction area using highlighter tool; the defined construction
414 area will show the grid view (with the unit of grid to be defined by the player).
415 (Click to) select the highlighted shipping container (or any constructional element/
416 object).
417 (Click to) select the grid in which the container (constructional object) will be positioned;
418 the object will be positioned or placed onto the grid.
419 Adjust the object’s position: move, rotate, and elevate up and down (with highlighted,
420 dotted 3D grid viewing showing).
421 During prototyping, the development team found it difficult to program an active
422 3D grid view for the gameplay mode using Unity 3D. In addition, the introduction
423 of a user-defined grid-view into the gameplay led to more difficult computerized
424 gaming assessment.
425 Coding game objects and developing the log structure are salient, interactive
426 processes that frame the 3D game world and in-game learning/performance assess-
427 ment. When coding major game objects, the team co-explored and outlined the
428 types, key properties, and associated basic functions of game objects based on (a)
429 an architectural building scenario script that describes and visualizes the exemplary
430 landscape of E-Rebuild (see Fig. 3.5) and (b) the targeted mathematical concepts
431 outlined in the aforementioned competency model. A constructional item object, for
432 example, can be outlined in the features of forms (e.g., cuboid, cylinder, triangular
433 prism, along with the basic functions of transforming, cutting, and scaling), physics
434 (e.g., friction, bounciness, density, and joint strength), materials (e.g., stone, wood,
435 brick), mass (e.g., solid or hollow), size (e.g., height, width, depth, radius, base,
436 angle), position, and occupancy (livable or not, empty to not). The key properties
437 and functions of each object class were then refined during the game programming
438 phase to better align with the interface of the Unity 3D development system.
439 Based on the defined object classes, the game programming sub-team then
440 drafted the game log structure that would drive the capturing of game performance
441 and provide the gameplay data to be mined for game-based learning assessment.
442 This log structure (see Fig. 3.6) was then reviewed and refined by the educational
443 measurement experts in the team to ensure that the actions and states logged were
444 in line with the task model, which was developed alongside the competency model,
445 and to provide sufficient high-quality data for a stealth assessment of the targeted
446 mathematical competency development.
447 Assessment of game play success was a major execution challenge that con-
448 fined the breadth of the previously defined game task evaluation and reward mecha-
449 nism (see Fig. 3.4). Detecting and scoring the player’s architectural artifact in the
450 game world was challenging mainly due to the conflicting demands of creativity
AU4 Fig. 3.5 Exemplary scenario script document for E-Rebuild. Historic Rural Southwest Scenario.
1 Adobe making yard with stacks of some adobes and mud pile, 2 Basic landscape and not many
trees, 3 Kiva with ladder, 4 Mountains in the background, 5 Orno ovens in yards, 6 Plaza with
rectilinear or square adobe houses around it, 7 Water of some kind…old-fashioned well and
pumps, 8 White church with graveyard, 9 Sheep herd. Suggested Uses of Elements for Design. 1
Adobe bricks (pick a size and this will be a great math problem), 2 Mud for mortar, 3 Sun for dry-
ing adobes
and feasibility. For example, a fully computerized evaluation of a 3D free-form 451

object based on the full set of architectural design and visual principles (e.g., sym- 452
metry, hierarchy, rhythm, transformation) involves developing and coding an 453
extremely extensive (if not interminable) list of measurable/observable features in 454
order to concretize the aesthetic representation, leading to endless programming 455
and debugging work of the task assessment. To confine the evaluation space, clear 456
and measurable design parameters were further delineated, prototyped, user tested, 457
and iteratively refined. The occupancy, position, size, and shape (or form), in rela- 458
tion to the limits of material, time, and living needs, were settled by the team as the 459
core design parameters for the evaluation of a built architectural structure. The 460
sub-team of game programming and the scientific computing experts then led the 461
discussion and the crafting of the specific formula for computing/scoring a build- 462
ing task (see Fig. 3.7). 463
Design pattern summary Design execution is a critical phase in which the team 464
reviews and gauges the integration and balance of feasibility, learnability, and play- 465
ability of each and every design proposition for the game mechanic, in-game learn- 466
ing, and assessment. Mainly via iterative prototyping in game engine, interdisciplinary 467
expert review, and infield user testing, the team selects design features that integrate 468
learning and play given limited time and resources and better plan the technical 469
details that frame the interactive game world, user interactions, and the interaction 470
performance logging and scoring. 471
Fig. 3.6 An example of game log structure draft

Fig. 3.7 A white-board sketch of the scoring of an architectural artifact
3.2.5 Structure and Navigation of Game Levels 472
Problem statement Designing the structure of game levels to create a responsive, 473
game-based learning path is another critical design inquiry for the E-Rebuild team. 474
For example, what should be the difficulty level for the initial game task/level? How 475
should a task, level, and multilevel episode be structured? Should the navigation 476
across the game levels be linear or nonlinear? When addressing these questions, one 477
could resort to multiple approaches of game structuring, including the segmentation 478
of an in-game adventure story, a gradual introduction of game mechanics, a scaffold 479
of mathematical learning and problem-solving, and/or an accumulation of perfor- 480
mance data to drive the training and test of an assessment model (e.g., a Bayesian 481
network). One might also argue for learner-selected, nonlinear navigation to support 482
learner autonomy versus a computer-controlled, sequenced navigation to guide 483
gaming and learning skill development. 484
Exploration of the design solution The team had tried to integrate a post- 485
quake-themed narrative, introduction of basic game rules and actions, and initia- 486
tion of mathematical conceptual learning when designing the initial game level, 487
as portrayed by the following design note: 488
489 The protagonist is trapped in a hole and needs to get out. All that is available are some
490 wooden blocks. The protagonist then needs to cross a river and all that is available are
491 some branches. (Initial problem/task used to train the actions of moving, rotating, and
492 stacking items.)
493 The protagonist comes across some victims who need a home built before a heavy rain
494 comes in. (Second problem—for mathematical problem-solving)
495 The victims can include four types of being, human adult, human child, animal adult,
496 and animal child, or animal above and under a specified weight. A human adult has a
497 standard requirement of space or cubic space (area or volume); the other three beings
498 will require different proportions of standard (cubic) space.
499 The site or the safe plane available for the home construction is limited (e.g., only 300
500 square feet, in certain irregular shape).
501 The protagonist comes across some victims who need a dam built before a river gets too
502 high and floods the village (Another problem for a future level, potentially).
503 The protagonist comes across some villagers who complained about their homes or
504 shelters as lacking light or in wrong direction toward wind, in comparison against their
505 neighbors (so in need of transformation such as rotation, reflection, or translation); some
506 others complained about the size of their homes, wanting them to be scaled up (other
507 problems for future levels, potentially).
508 Based on the above design note, we then developed a mock-up of the initial game
509 level that encompassed the prototypes of the game world (e.g., earthquake simula-
510 tion, blocks and containers as the building items, and victim characters), basic gam-
511 ing actions (e.g., collecting and maneuvering items to build, allocating victims to
512 the built structure), constraints/rules (e.g., time, materials, design parameters, and
513 living space requirements), user interface (e.g., a help panel, feedback, a measure-
514 ment tool), and evaluation rules of the structure built (e.g., checking whether the
515 size/location/direction replicates the pre-quake model to 90%). Specific mathemati-
516 cal context problems embedded within the game levels were drafted by the mathe-
517 matical education expert in the team based on a collection of state and common core
518 test items for middle-school students. An exemplary draft/sketch of a game-based
519 math problem along with the design notes are provided below (Figs. 3.8 and 3.9):
520 Building a shelter via shipping containers to accommodate the victims saved:
521 16x10 + 8π square meters open space for potential construction, 67 people (40 children,
522 20 adults, 7 pets), 16 small containers (6 x 3 x3), and 8 big containers (12 x3 x3)
523 requiring 5 meters higher than the base for living space
524 Minimum space for pet, child, and adult: 1:2:4
525 Maximum space for pet, child, and adult (meaning numbers bigger than this maximum
526 will not influence living being’s happiness any more): 2:4:8
527 The mock-up was iteratively refined through user testing with a small group of
528 middle-school students in a local school’s after-school program. It then served as
529 the initial archetype for the development of additional game levels.
530 Structuring and navigation To plan and explore the structuring and navigation of
531 game levels, the team has referred to instructional design strategies (particularly
532 elaboration theory, Reigeluth, 1992), prior work on flow experience in gaming (e.g.,
533 Csikszentmihályi, 1990; Kiili, 2005), and discussions of evidence accumulation
534 across tasks in the evidence-centered design approach (Mislevy, Almond, & Lukas,
535 2003). A conceptual framework on how to sequence and navigate levels based on
Fig. 3.8 A design sketch of the initial game level
the embedded math competencies, the highlighted architectural design scenarios, 536
and the difficulty index of tasks, was then drafted (see Fig. 3.10). We then proposed 537
that the player could proceed across levels by (a) exploring tasks of the same archi- 538
tectural design scenario and same subset of math topics but varied difficulty levels 539
and then advancing to a different scenario with different subsets of math topics (as 540
portrayed by the blue line in Fig. 3.10) or (b) exploring varied architectural design 541
scenarios along with different sets of math topics at a similar difficulty level and 542
then advancing to more challenging tasks (as portrayed by the red line in Fig. 3.10). 543
Path a or b can be set by the computer as a fixed, linear sequence for all players. In 544
comparison, the game can allow the players to self-choose a path, to skip certain 545
levels, or even to explore any levels without a sequence/path. 546
The team was indecisive as to which path to employ. Path a appeared to be more 547
aligned with E-Rebuild level development sequence and hence was executed first in 548
our initial infield feasibility studies with 66 6th–7th graders in their science or math 549
classes. The infield observation indicated that players differed in their gameplay 550
progress and naturally shared/compared their gaming experiences. While some 551
remained stuck in Scenario 1, others had proceeded to Scenario 2 or even Scenario 552
3. The former group demonstrated obvious frustration when their peers showed off 553
the new game landscapes and items. Besides, not all students managed to complete 554
all scenarios within the study sessions and hence process all embedded mathemati- 555
cal concepts, which made it challenging to accumulate sufficient evidence or data 556
across tasks for the game-based math competency/learning assessment. It also 557
reduced the potential to provide an equivalent access to game-based mathematical 558
Fig. 3.9 Computerized prototyping of the first E-Rebuild level
559 learning content. Given these infield findings, we then switched to Path b in the later
560 infield implementation studies and found that the aforementioned issues all got
561 resolved. Another infield finding from our design experiments with the game level
562 structuring and navigation was that middle-school students were not adaptive to the
563 open-ended task structure and frequently requested help with chunking and sequenc-
564 ing steps of the design problem-solving in a game level (Ke et al., 2017). Even
565 though some E-Rebuild team members had argued for the sense of autonomy
566 afforded by a nonlinear navigation with the player choice, our experimental findings,
3.3 Constructive Conflict, Group Synergy, and Leadership 71
Fig. 3.10 Game level sequencing and its integration of math competencies and design scenarios
consistent with prior research (Kim, Almond, & Shute, 2016), implied that nonlinear 567
gameplay did not produce better game engagement or task performance but could 568
negatively impact the evidence accumulation for game-based learning assessment. 569
Therefore, a linear gaming structure was adopted. 570
Design pattern summary In a learning game that aims to incorporate content 571
learning, authentic problem-solving, and stealth learning assessment without inter- 572
rupting the game flow, the structure and navigation of game levels should be 573
designed to scaffold play, learning, and assessment simultaneously. Specifically, the 574
initial game levels should train the player on core game mechanics and frame the 575
backdrop mission or theme narrative that embed, represent, and contextualize 576
domain-specific concepts or problem-solving. Game structuring, encompassing the 577
structuring of targeted competencies, scenarios, and tasks variant in difficulty, 578
should scaffold conceptual or skill development as well as evidence accumulation 579
for learning assessment. Game-based navigation or learning path should be selected 580
and dynamically adjusted based on learners’ response and performance during the 581
infield testing. 582
Aside from the aforementioned salient design processes and patterns, constructive 584
conflicts, group synergy, and leadership were three emerging, prominent themes 585
found to advance segmented multidisciplinary perspectives and skills toward coher- 586
ent interdisciplinary design. They developed given purposeful preparations that 587
took some effort. These preparations involved (a) confronting and discussing issues 588
589 of scope and definitions to frame a common design goal or space, (b) using a com-
590 mon set of design aids to identify linking points between disciplinary propositions
591 and activate group synergy in the solution exploration, and (c) building participative
592 and transformational leadership to reduce deconstructive conflict and create effi-
593 cient decision-making.
594 3.3.1 Framing a Common Design Goal and Space
595 Team members of different disciplines initially proposed interpretations of the

596 design goal or the scope of design space for E-Rebuild, as implied by the following
597 meeting note that recorded team members’ perspectives on the most unique feature
598 of E-Rebuild.
599 • Mini-me (embodied via a 1.5 meter middle-school student figure) performing
600 various architectural design/building activities.
601 • Integrating math history into the gameplay, for example:
602 An AI agent (history agent) has time-traveling power and can present not only
603 the views of the site/architectures before an earthquake but also the history of
604 ancient construction tools (e.g., perpendicular lines in the coordinate plane)
605 or practices (e.g., proportion reasoning when measurement units are not stan-
606 dardized or existing).
607 • 3D object maneuvering and transformation
608 • Development and validation of game-based stealth assessment of learning.
609 The above perspectives were presented by the scholars from the disciplines of
610 architecture, mathematics education, scientific computing, and educational mea-
611 surement, respectively. Such a diverse range of design priority propositions were
612 found to expand the scope and create a segmented agenda when framing the design
613 space. Thus, our initial design meetings mainly focused on consolidating design
614 goal interpretations.
615 Moreover, an interdisciplinary design team has different customs of classifying or
616 naming design concepts, which made the design discussion and brainstorming at the
617 initial design meetings especially challenging. For example, scholars of learning sys-
618 tems design, architecture, and scientific computing all had variant usages for terms.
619 Some examples include “blueprint” versus “floor plan” or “site plan,” “God’s view”
620 versus “plan view” or “elevation view,” and game “level, episode, or chapter,” which
621 led to an initial discussion on the design of gaming perspective and the game world
622 projection confusing and clumsy. Documents explicitly defining architectural vocab-
623 ulary, disciplinary concepts, and gamer terms were hence drafted to assist interdisci-
624 plinary design communication, as the following example illustrates (Table 3.1).
Table 3.1 An exemplary sketch specifying gaming and scientific computing vocabulary t1.1
Building items (gaming) Objects (scientific computing/coding) t1.2

Forms Forms: cuboid, cylinder, sphere, triangular prism, t1.3
cone, pyramid t1.4
Transformation, cutting, scaling t1.5
Materials (e.g., stone, wood, brick, and Physics: friction, bounciness, density, joint strength t1.6
steel, glass?) t1.7
Solid/hollow Mass (Can hollow objects be cut? Center of mass?) t1.8
Size Height, width, depth, radius, base, angle t1.9
Positioning t1.10
Livable space Occupancy t1.11
3.3.2 Using Common Design Aids 625
Exemplary design cases in architectural and learning games were used as the 626
“seeds” or the design aids to help the team identify linking points between disciplin- 627
ary propositions and activate group synergy in the solution exploration. The follow- 628
ing project meeting note highlighted the usage of design examples as the starting 629
point for the team to identify the linking points between disciplinary propositions 630
and activate group synergy in the solution exploration: 631
9/28 632
We have decided that before next meeting, we will further explore gaming tasks and core 633
mechanics of E-Rebuild by reviewing the Bridge! Construction game, Minecraft, and 634
Google SketchUp. The following is a suggested list of design questions to be considered/ 635
addressed during the review. 636
What are the major game mechanics? Is it given resources and design constraints, 637
build or compose architecture? 638
What is the potential of mixing survival mode (speeded construction task) and creative 639
mode (simulation mode in which one can test the construction product with the 640
earthquake tool)? 641
Game actions: Moving, stacking, cutting, and scaling? Model sketching within the 642
game? 643
Rule: Limit of constructional items as a constraint of design? Reward the level pass 644
with more inventory items? 645
Will a building block (e.g., a cube or shipping container) be given by the system? How 646
and why should the system give the player with shipping containers? Any story 647
element? 648
What game rewarding system? 649
How will the tools be given to players, one at each game level, or once for all at 650
beginning? 651
How will players learn to play? 652
We have decided that each teammate should present a design sketch or a written 653
document to address the above questions and then have all thoughts synthesized during 654
next meeting. 655
656 3.3.3 Building Participative and Transformational Leadership
657 As the above design note illustrates, the team of E-Rebuild resorted to participative
658 design practice that values the input of every team member and holds everyone
659 responsible for design brainstorming. A prerequisite of such a design style was the
660 lengthy discussion and congregation of each and every perspective in the project
661 meetings, extending the duration of the group meeting. Frequently, a common syn-
662 thesis could not be reached, and the meaningful negotiation was abruptly stopped
663 when a meeting adjourned. To increase the efficiency of decision-making and ensure
664 the design discussion would be thorough, we divided the whole team into groups
665 based on major design responsibilities, with each group led by a disciplinary expert.
666 We also separated the design meetings into two sections, the whole-project meeting
667 and the sub-team design meeting. At the former ones, we focused on the big picture,
668 sharing the reports of every sub-team and conducting high levels of communication/
669 decision-making to accomplish design goals. Smaller and detail tasks were dele-
670 gated to each sub-team who would hold separate sub-team design meetings to fully
671 discuss and explore the solutions to discipline-specific design inquiries. These solu-
672 tions would be brought back to the whole-project meetings for review and critique.
673 In general, we observed that such a mixture of participative and transformational
674 leadership (Burns, 1998) aided in efficient and constructive design decision-making
675 in an interdisciplinary team.
676 References
677 Burns, J. M. (1998). Transactional and transforming leadership. In G. R. Hickman (Ed.), Leading
678 organizations: Perspectives for a new era (pp. 133–134). Thousand Oaks, CA: SAGE.
679 Csikszentmihályi, M. (1990). Flow: The psychology of optimal experience. New York: Harper &
680 Row.
681 Garris, R., Ahlers, R., & Driskell, J. E. (2002). Games, motivation, and learning: A research and
682 practice model. Simulation & Gaming, 33(4), 441.
683 Habgood, M. P. J., & Ainsworth, S. E. (2011). Motivating children to learn effectively: Exploring
684 the value of intrinsic integration in educational games. The Journal of the Learning Sciences,
685 20(2), 169–206.
686 Ke, F. (2016). Designing and integrating purposeful learning in game play: A systematic review.
687 Educational Technology, Research and Development, 64(2), 219–244.
688 Ke, F., Shute, V., Clark, K., Erlebacher, G., Smith, D., Fazian, P., Lee, S., & Xu, X. (2017, April).
689 Math learning through game-based architectural design and building. Paper to be presented at
690 2017 American Educational Research Association Annual Meeting, San Antonio, TX.
691 Kiili, K. (2005). Digital game-based learning: Towards an experiential gaming model. The Internet
692 and Higher Education, 8(1), 13–24.
693 Kim, Y. J., Almond, R. G., & Shute, V. J. (2016). Applying evidence-centered design for the devel-
694 opment of game-based assessments in physics playground. International Journal of Testing,
695 16(2), 142–163.
References 75
Mislevy, R. J., Almond, R. G., & Lukas, J. F. (2003). A brief introduction to evidence-centered 696
design. ETS Research Report Series, 2003(1). Retrieved from https://www.ets.org/Media/ 697
Research/pdf/RR-03-16.pdf 698
AU5 Moustakas, C. (1994). Phenomenological research methods. Sage. 699
Pratt, D. D., Winters, N., Cerulli, M., & Leemkuil, H. (2009). A patterns approach to connecting 700
the design and deployment of mathematical games and simulations. In Technology-enhanced 701
learning (pp. 215–232). Dordrecht, Netherlands: Springer. 702
Reigeluth, C. (1992). Elaborating the elaboration theory. Educational Technology Research & 703
Development, 40(3), 80–86. 704
Winters, N., & Mor, Y. (2008). IDR: A participatory methodology for interdisciplinary design in 705
technology enhanced learning. Computers & Education, 50(2), 579–600. 706
Author Queries

AU2 Please confirm the inserted citation for Figs. 3.1 and 3.4.
AU3 Please provide better quality artwork for Figs. 3.1, 3.3, and
3.5.
AU4 Please confirm the caption for the Fig. 3.5.
AU5 Please provide publisher location for Moustakas (1994).
Chapter 4 1
Design of Gameplay for Learning 2
Abstract A common skepticism about educational games is that learning and play 3
are frequently not well integrated—the skill or content to be used and learned lacks 4
a semantic or meaningful relation with the fantasy and challenge elements and can 5
be easily swapped without influencing gameplay. In this chapter, we describe and 6
analyze design challenges associated with the core components of gameplay— 7
game mechanics and the narrative scheme for the learning purpose—and review the 8
gameplay design propositions and infield test findings of E-Rebuild. Via a retro- 9
spective investigation of design features and strategies in terms of learnability and 10
playability (i.e., capability of activating knowledge-based cognitive performance 11
without interrupting gameplay), this chapter aims to report and discuss how domain- 12
specific learning is integrated in and activated by core game actions, rules, game 13
objects, and the game world design. 14
4.1 Introduction 15
AU1 Gaming is an organized play structured by a set of rules and an obstacle-tackling 16

goal (Klopfer, Osterweil, & Salen, 2009; Schell, 2014; Suits, 1978). Learning is an 17
innate element of gaming, in that players interact with a game to learn rules and play 18
strategies and then adapt and improve play skills to make progress in the game 19
(Lindley & Sennersten, 2008). Yet learning in games varies in the degree to which 20
the acquired skills and strategies emphasize “domain-specific knowledge” (Ke, 21
2016; Tricot & Sweller, 2014). The higher-degree domain-specific knowledge is 22
needed, the more careful design effort is required for the attainment of an integral 23
relationship between gaming and learning (Habgood & Ainsworth, 2011; Kafai, 24
1995; Malone & Lepper, 1987; Torbeyns, Lehtinen, & Elen, 2015). A common 25
skepticism about educational games (and, hence, a relevant design issue) is that 26
learning and play are frequently not well integrated—the skill or content to be used 27
and learned lacks a semantic or meaningful relation with the fantasy and challenge 28
elements and can be easily swapped without influencing gameplay. As a result, 29
players may be distracted by the play part and not achieve the learning goals, or they 30

https://doi.org/10.1007/978-3-030-04339-1_4
78 4 Design of Gameplay for Learning
31 may become disengaged because learning elements may corrupt what is enjoyable
32 about games (Garris, Ahlers, & Driskell, 2002; Miller, Lehman, & Koedinger,
33 1999). For example, in E-Rebuild we could design a “trading” challenge that
34 requires the player to complete a screen of math questions to obtain the corresponding
35 number of game tokens and then use the tokens earned to purchase items needed.
36 Such a game mechanics (i.e., the question-answering point system) could be reused
37 for another challenge (e.g., earning time credits to buy more time for a “building”
38 action) only with the content or math topics of the questions swapped as needed. An
39 extrinsic integration of content in gameplay, as the example illustrates, appears to be
40 a rapid and all-purpose gameplay design strategy. But it is, as prior research
41 suggested, deficient in reinforcing game-based engagement and won’t likely
42 promote desirable active and deep learning, especially for learners who lack the
43 competency of and positive disposition toward the subject matter (Ke, Xie, & Xie,
44 2016; Richards, Stebbins, & Moellering, 2013).
45 In this chapter, we describe and analyze design challenges associated with the
46 core components of gameplay—game mechanics and the narrative scheme for the
47 learning purpose—and review the gameplay design propositions and infield test
48 findings of E-Rebuild. Via a retrospective investigation of design features and strate-
49 gies in terms of learnability and playability (i.e., capability of activating knowledge-
50 based cognitive performance without interrupting gameplay), this chapter aims to
51 report and discuss how domain-specific learning is integrated in and activated by
52 core game actions, rules, game objects, and the game world design.
53 4.2 C
onceptualization of Game Mechanics, Narrative,
54 and Learning Integration
55 Endorsing the perspectives of the previous game studies, we have designed and
56 examined the construct of gameplay in two layers: the “ludus” or game mechanics
57 layer that involves rules and actions and the narrative layer that comprises the set-
58 ting (or scenario), backdrop mission, and game objects (Ang, 2006; Frasca, 1999).
59 It is commonly believed that gameplay lies in the meaningful interplay between the
60 two layers, though whether game design is more the design of experience (Salen &
61 Zimmerman, 2004) or a narrative architecture (Jenkins, 2002) is still inconclusive.
62 According to Järvinen (2008) and Sicart (2008), the term game mechanics refers
63 to an activity structure consisting of rules and play actions. Rules are designed to
64 determine the conduct and standard for both play behaviors and the winning/losing
65 state. These rules lead to the creation of player strategies and actions with which the
66 player can interact with game elements to “influence the game state towards the
67 attainment of a goal” (Järvinen, 2008, p. 255). The game mechanics hence is “a
68 compound activity composed of a suite of actions” that players, abiding by the
69 rules, recurrently perform and directly apply to achieve the goal state (Salen &
70 Zimmerman, 2004, p. 316; Sicart, 2008).
4.3 Learning-Play Integration in E-Rebuild 79
Not all games tell a story, and hence the narrative layer is not a defining feature 71
of games (Ke, 2016). But many games do have narrative aspirations, or at least tap 72
into the player’s memory of previous narrative experiences (Jenkins, 2002). 73
According to Jenkins (2002), narrative can be integrated into a game as a broadly 74
defined goal, such as a backdrop plot or mission, (b) a localized incident or plot 75
developed in game level(s), and/or (c) an open-ended game world that allows play- 76
ers to define their own stories via authoring- or construction-based play. Notably, 77
a game can create an immersive narrative experience and convey storytelling via 78
the “spatiality” of the game world, in which the narrative element is infused into a 79
space that a player navigates through and interacts with. In E-Rebuild, we have 80
explored the aforementioned three approaches of the narrative layer during the 81
design process and in particular focused on examining the feasibility and effec- 82
tiveness of representing a mathematical story problem via the spatial narrative of 83
the game world. 84
In a recent literature review of domain-specific content learning via gameplay 85
(Ke, 2016), the typical processes of integrating learning in game actions are 86
described as representation, simulation, and contextualization. Specifically, repre- 87
sentation involves designing game objects as external representations of the tar- 88
geted concepts and interactions with game objects as the conceptual exploration or 89
application processes. Simulation is achieved via designing the game world as a 90
scientific problem or system and hence the game actions as an iterative process of 91
problem-solving and discovery learning. Contextualization, conveyed via a back- 92
drop mission, a plot, game characters, and/or meaningful scenarios, is then employed 93
to increase the pertinence and fascination of content representation and problem- 94
solving simulation for players. 95
The design of E-Rebuild has been inspired by previous sandbox or architecture 97

game examples (e.g., Minecraft, SimCity), 3D modeling tools (e.g., Google 98
SketchUp), and the previous architectural design education programs for children 99
(e.g., Architecture and Children Pilot Curriculum; Taylor, 2009). Different from 100
and extending these design stimuli, E-Rebuild needs to align and combine construc- 101
tion-/authoring-themed gameplay, accessible 3D design, and interaction interface, 102
as well as architectural design missions and setting, to facilitate mathematical learn- 103
ing and problem-solving for middle-school students. Designing core game mechan- 104
ics that manage to embody such an interdisciplinary design goal is critical and 105
challenging. What should be the nature or foundational “act” of the interdisciplin- 106
ary, learning-constructive gameplay? What corresponding game actions and rules 107
are found to reinforce the game’s learnability and playability? How will backdrop 108
and virtual narrative environment contextualize or “legitimize” the target learning 109
and play actions? 110
111 The aforementioned questions underlay our design inquiry on the core game
112 mechanics of E-Rebuild. We conducted a retrospective and thematic analysis with
113 the design notes, meeting records, design talk among teammates at meetings, and
114 the observation notes of players’ actions and reactions when test-playing the proto-
115 typed E-Rebuild game levels. The analyses focused on delineating salient design
116 propositions and beliefs governing the core game mechanics and narrative scheme,
117 refined perceptions and design moves during infield testing, and the identified, func-
118 tional design heuristics that promote in-game learning actions without interrupting
119 the flow of play. These design research findings, with support of citations and exam-
120 ples, are presented below in a succession of themes that collectively constitute the
121 thinking process and core sectors of the game mechanics and narrative design
122 process.
123 4.3.1 Setting Genre and Nature of the Core Gameplay
124 There were two intertwining threads of design effort in the initial inquiry of core
125 gameplay design in E-Rebuild—searching the genre of building-themed gameplay
126 and investigating the nature of building-relevant mathematical experience. The for-
127 mer thread of design effort highlighted the element of play by addressing the fol-
128 lowing questions: “What will be building-themed play? What is the core action of
129 play? What will be the major obstacle and goal for such a play act?” The latter
130 thread of design, differently, focused on the learning experience by defining the core
131 components of a salient or desirable mathematical experience or thinking process.
132 The intent was to seek the salient and unified actions existing between architectural
133 design and mathematical knowledge application and gauge how and why these
134 interdisciplinary actions could be challenging while personally meaningful for stu-
135 dent players to be “fun.”
136 Genre or type of building-themed gameplay Why is building fun? was the head-
137 line title of the first design meeting memo of the E-Rebuild project, exemplifying
138 the predominant and early design effort contributed to exploring the potential genres
139 of serious play enabled by the processes of planning, designing, and constructing
140 buildings and other physical structures. Such an exploration was activated via (a) an
141 inspection of toys, games, simulations, and other hypermedia examples highlight-
142 ing the theme of building, especially survival rebuilding activities, and (b) a reflec-
143 tive design conversation among the design team members in relation to our own
144 experiences and preferences of building-themed gameplay.
145 The design team members of E-Rebuild have, individually and collectively,
146 sought, reviewed, and test-played a collection of architectural construction- or
147 building-oriented toys, games, simulations, and other hypermedia applications. The
148 assorted prior examples of building-themed serious play demonstrated various
149 modes of play, such as block stacking puzzle (e.g., https://play.google.com/store/
150 apps/details?id=com.ketchapp.stack), block building in the real or virtual world
(e.g., Magna-Tiles, Minecraft), adventure (e.g., Raw Danger!, Camelot Unchained), 151
strategy game (e.g., http://www.stopdisastersgame.org/en/home.html), and physics 152
simulation (e.g., Bridge Constructor, Bridge Construction Simulator). These exam- 153
ples of building-themed gameplay could evolve around a designer-defined game 154
goal or the preset criteria for game success, such as height and balance in block 155
stacking, sturdiness of the construction, survival, and/or people and property protec- 156
tion against a disaster. It may also depict an open-ended, self-determined goal of 157
play, such as creative play via block building in a sandbox game. Core player or 158
user actions include block stacking (information or object) collection, and decision- 159
making, with real-time physics simulation activated, elective, or deactivated during 160
those actions. Reflecting a focus on an adventure or a construction mode of play, the 161
core gaming action will adopt a plan view or an elevation view, whether in first- or 162
third-person perspective. Ultimately, the obstacles to be resolved to achieve the play 163
success are typically the challenge posed by the properties of the building materials 164
in relation to the construction goal (e.g., the weight, shape, structure, variety, and/or 165
scarcity of construction materials in a block stacking puzzle or a bridge construction 166
game), time limitation, and insufficiency of prerequisite knowledge or skills of a 167
construction task (e.g., engineering concepts in Minecraft; physics in Bridge 168
Constructor; science, economics, and global issues in Stop Disasters; or generic 169
problem-solving skill in Raw Danger!). 170
The design case review and test-play notes of the design team indicated variabil- 171
ity among the team members in terms of the play preference, discipline knowledge, 172
and hence the thought processes for design. For example, regarding two exemplary 173
construction- or building-themed games, a game designer commented, “I think 174
what we can learn from these two games, are the game actions and the potential 175
design constraints we can have: Given a landscape or terrain, given the potential risk 176
of future hazards and limited budget and materials, where and what kind of post- 177
quake architectures should we invest on and build?” The comment carried the game 178
design language (e.g., “game actions” and “constraints”) and a designer perspective 179
governing the common core game mechanics in the game examples. Yet two team 180
members, a mathematician and an educational measurement specialist, made differ- 181
ent annotations (provided below). These annotations highlighted the player percep- 182
tion, indicating different player preferences. 183
https://eduweb.com/portfolio/bridgetoclassroom/engineeringfull.html 184
http://www.stopdisastersgame.org/en/home.html 185
Cool games. I just looked them over quickly, and they seem interesting! 186
I played the bridge game and it was boring. The other earthquake game was a bit more 187
engaging, but not by much. We have a chance to build a more compelling game for this 188
niche! 189
The second commenter then expressed her preference of adventure games (e.g., 190
Raw Danger!) that got concurred by another team member who was also a fan of 191
sandbox games like Minecraft. 192
Indeed, play preferences tended to mediate how individual team members framed 193
the role of building or constructing in the play experience. Some perceived the task 194
of construction or building by itself as a fun process, thus framing the core gameplay 195
196 as the simulation of building actions and replicating architectural design constraints
197 and specification as the obstacles and goal state. Others considered building mainly
198 a backdrop mission or a meaningful context that legitimizes the gameplay of cre-
199 ative problem-solving and adventure. Another individual characteristic that medi-
200 ated the thought processes for building-themed gameplay design is disciplinary
201 knowledge or epistemology. The architecture educator, for example, insistently
202 emphasized the basic principles of design (e.g., balance in visual distribution,
203 rhythm, unity) and creativity as core goal states or success evaluation criteria. Yet
204 game designers and programmers in the team preferred measurable objectives (e.g.,
205 time, thriftiness in material usage, structural soundness, and building specifications
206 of a structure) and expressed concern as to whether and how some principles and
207 elements of design can be defined and specified quantitatively for computerized in-
208 game evaluation. Certain team members also worried that giving players unlimited
209 materials and seeing what they create would eliminate constraints, thus reducing the
210 challenge and hence fun in the game.
211 After multiple rounds of reflective and active conversations with the design cases
212 and teammates on the type of building-themed gameplay, we decided that our proto-
213 type of the core game mechanics would try to integrate different design perspectives,
214 which would be infield tested by the learners and then refined based on the testing
215 data and learners’ feedback. Our initial prototype of building-themed gameplay,
216 therefore, encompassed the following elements: (a) shifting between a first-person,
217 adventure mode (with collection, site measurement, and space allocation as sector-
218 specific game actions) and a third-person, building mode (with trading and building
219 as core game actions, Fig. 4.1); (b) simulating physics (e.g., earthquake waves) to
220 test the structural soundness of a building artifact; (c) enabling both plan and eleva-
221 tion view in the building mode (Fig. 4.2); (d) depicting building obstacles as the
222 variety of landscapes, time, and material limit; (e) defining/evaluating the goal stated
223 mainly via time, materials usage, structural soundness (e.g., not collapsing given
224 variant earthquake waves), and preset building specifications (e.g., size, location, and
225 direction aligned with a pre-quake house model); and (f) letting peers and potentially
226 the teacher rate the aesthetic design elements and originality of the structure outside
227 the game during post-game debriefing. This prototype later got iteratively reviewed,
228 user tested, and refined considerably, as described in a later section.
229 Building-relevant mathematical experience Another concurrent thread of design
230 inquiry on core gameplay in E-Rebuild was to investigate the salient, common fea-
231 tures of both a mathematical experience and an architectural design/construction
232 experience, or What makes a good building task or activity that creates engaging
233 and effective mathematical experiences for participants? Exploring building-rele-
234 vant mathematical experience is critical for an intrinsic integration of learning and
235 building in gameplay. The exploration process evolved around the interviewing of
236 disciplinary experts, the literature and design case review, and an inspection of
237 exemplary design education curricula (e.g., Architecture and Children Pilot
238 Curriculum; Taylor, 2009) that promote the dynamic architectural and mathematical
239 experiences for children.
Fig. 4.1 Adventure and building modes of play
The design discussions on the joint features of architectural and mathematical 240
experiences involved the analysis of core actions, basic resources or materials, fre- 241
quently used tools, requisite knowledge and skills, and a set of archetypal practices 242
typifying the aforementioned elements. For instance, the following game design note 243
carried initial discussions on the potential conjoint architecture-math experience 244
elements. 245
Fig. 4.2 Plan and elevation views
246 Tools
247 • Measuring/layout tools: rule, square, protractor, etc.
248 • Sketching tools: brush.
249 • Construction tools: e.g., cutting, transportation, elevation tools, etc.
250 Basic materials
251 • Building unit: line, plane, form, or prefabricated block.
252 • Shape: cube, arch, cylinder, etc.
253 • Texture: stone, brick, etc.
254 Actions
255 • Navigate and explore (i.e., measure, collect data, and analyze data).
256 • Draw or sketch via extended devices and apps—mixed reality.
257 • Build: select, drag, drop, and customize.
258 • Test (e.g., with simulated earthquake waves) and predict, decide, or reflect.
259 • Collect, trade, and/or manage resources (e.g., time, space, material cost).
260 Exemplary Practices
261 • Develop architectural plans and a model for a “mouse” house: Bubble dia-
262 gram –> floor plan (a plan view) –> elevation drawing (2D drawings that show
263 the outside walls of a building, e.g., the front side with door) –> perspective
264 drawing (a house that is three-dimensional looking) –> model construction.
265 • Related task (p. 8 in Taylor, 2009): Designing a vegetarian fast-food restaurant
266 (eating and preparation areas, delivery space storage, restrooms, parking,
267 etc.—spaces only, not equipment).
268 • Predesign exercise or puzzle: decorate a shoe
269 • Structure in architecture: arch, triangle, or asymmetrical vs. balanced.
270 • Related task: The picnic shelters must use a space frame structure for their roofs.
271 Each shelter will have four picnic tables, and the space frame will rest on at least
272 four columns.
• Predesign exercise or puzzle: Toothpick puzzles 273

• Form in architecture, involving measurement, data interpretation, scale, and 274
proportion. 275
• Related task: Use organic structures found in nature in a “fantasy” architectural 276
structure for the “moon.” 277
• Predesign exercise or puzzle: Scavenger hunt for basic geometric shapes and 278
forms in nature (e.g., crystal, plants’ spirals, spider’s net) (p. 31). 279
• Related task: Design a “people’s” bridge across the canyon to build a jogging 280
and bike path. This bridge should be as beautiful as the landscape that surrounds 281
it and should appear to gracefully and sturdily cross the canyon (p. 35). 282
• Predesign puzzle: Designing and modeling natural bridges—a fallen tree as a 283
beam bridge, intertwining vines as suspension chains or bridge, and a rockfall 284
and later the water eroded some rocks away forming a simple arch over the 285
stream (i.e., the arch bridge) (p. 33) 286
• Visual design: Super wall graphics using concrete block or block grid pattern 287
enlargement method (scale, proportion, measurement, introduction of area and 288
perimeter) (p. 38). 289
• Related task: Drawing and painting a supergraphics for your village’s school, 290
which will represent your community. 291
• Predesign exercise or puzzle: Painting an existing graphic projected onto a brick 292
wall. Mixing color and texture 293
• Creative, imaginative design: My favorite place. 294
• Related task: Design and model an “ideal classroom” using a scale, and then 295
construct it. 296
• Similarities between body and building systems (membrane-skin, mechanical 297
systems-muscular system, heating/cooling-circulatory system, electrical system- 298
nervous system, structural system-skeleton, waste disposal system-digestive sys- 299
tem, ventilation-respiratory system, mechanical devices that lift and push 300
things-muscular system). 301
• Related task: Transformer “You are architecture.” 302
• Landscape (p. 45)—collaboratively make a section drawing from a topographic 303
map of a site, schematic drawing of individual designs, and create models and 304
the final working architecture. 305
• Related task: Redesign a creative play park for the surrounding community and 306
the school. 307
• City planning—collaboratively plan, model, and build transportation, residen- 308
tial areas, business district, recreation and open spaces, and the capitol complex 309
(p. 50) (data collection and interpretation, in addition to all prior knowledge). 310
• Predesign exercise or puzzle: mapping and touring. 311
As illustrated in the above design notes, our initial exploration of building-rele- 312
vant mathematical experience has pinpointed a short list of core actions—site analy- 313
sis (surveying), sketching, building (design) experimentation, and resource 314
management. Yet there is still much ambiguity and disagreement existing in the 315
team discussions: Will the experience or actions highlight design/sketching (e.g., 316
317 drawing and visual pattern design) or building/modeling (e.g., the composition and
318 decomposition of the forms and shapes) or both? What should be the granularity
319 level of the building action? What are the various forms the building units can
320 assume? What and how will the tools be integrated into the game setting?
321 The design discussion on the role of design/sketching versus building/model-
322 ing in the gameplay experience was actually a comparative analysis on the two
323 prospective game mechanics in enabling and legitimizing targeted mathematical
324 practices. In particular, we discussed whether and via what enabling acts or moves
325 the two mechanics would embody the multimodal conceptualization and applica-
326 tion of the targeted math knowledge (e.g., angle measure, area, and volume of
327 geometric figures). Via expert interviews and exemplary task performance obser-
328 vation along with task analyses, we found that the game mechanics of building or
329 modeling tends to optimally align with the targeted learning actions (Table 4.1).
330 We then wondered about the applicable user input interface for each action and
331 the feasibility of in-game tracking and assessing of design or building moves.
332 Design, especially sketching, as the game mechanics would entail high-precision
333 touch screen interaction (e.g., tablets or laptops with stylus); the complexity
334 involved in architectural drawing and creative design also composed challenges
335 to real-time, in-game behavior tracking and design artifact evaluation. As such,
336 gameplay in E-Rebuild portrayed more building/modeling actions than design/
337 sketching actions.
338 The granularity level in the basic unit to be maneuvered during the building/
339 modeling actions would vary. Specifically, using a “block” approach, the game
340 could provide prepackaged structures (e.g., a shipping container to be remodeled as
341 a house) as building blocks to be composed, decomposed, or modified to create vari-
342 ant structures. Alternatively, a player would build using the basic architectural forms
343 (e.g., bricks, doors, windows) to compose higher-level elements (e.g., walls, floors,
344 and roofs), then a house or bridge, then a neighborhood, and so forth. During design
345 discussions, the team perceived that both approaches have advantages. The design
346 hypotheses were that the previous approach would enable a novice player to
347 construct a complicated structure efficiently while focusing his effort on planning
348 before building. The latter, though demanding a higher level of design knowledge
t1.1 Table 4.1 Actions of learning and building

t1.2 Learning actions Building actions
t1.3 Identification and representation (e.g., of 2D/3D Collection and trading (of construction
t1.4 geometric figures) elements)
t1.5 Measurement and calculation (e.g., of length, angle, Construction site survey and plan
t1.6 circumference, area, and volume)
t1.7 Creation Building—positioning, rotating,
t1.8 stacking, joining, copy-pasting
t1.9 Analysis Allocation of space, time, and other
t1.10 limited resources
t1.11 Evaluation Design evaluation: fulfilling needs,
t1.12 efficiency in cost and time
and technical fluency, would situate building units in a variety of forms in line with 349
the system of geometric shapes to be learned. Hence both approaches were used and 350
integrated across game episodes and tasks in E-Rebuild. 351
To implement building-relevant gameplay, the user input interface is another 352
critical design element. The interface should enable core game-based learning 353
actions while keeping players in control. In other words, it should facilitate action- 354
based mathematical conceptualization or knowledge application beyond affording 355
simplicity and efficiency. To explore appropriate user input interfaces for core game 356
actions, we designed multiple papers, toys, and computerized prototypes, iteratively 357
tested them with expert and novice gameplayers, and evaluated the interaction expe- 358
riences to select the interface and interaction tools that are not only user-friendly but 359
also learning constructive. Detailed discussions of the user input interface or tool 360
design are provided in a later section. 361
4.3.2 D
esigning Learning-Play Integration in Game Actions 362
and Rules 363
During the past 4 project years, we have iteratively designed, tested, and refined 364
more than 30 prototypes of E-Rebuild. Learnability and playability are two key 365
objectives driving the iterative design tests and refinements. During iterative testing, 366
we have observed a continuum of gameplay engagement framed by the designed 367
game actions and rules—from play engagement (i.e., game-based play irrelevant to 368
the designed tasks), task engagement, and cognitive engagement to content engage- 369
ment, being incremental in demonstrating commitment to mathematical thinking, 370
knowledge application, and problem-solving. After iterative refinement and elimi- 371
nation, the resulting game actions and rules have the following core feature or func- 372
tionality: necessitating cognitive and content engagement within gameplay actions 373
and rewarding mathematical thinking as the most warranted gameplay strategy. 374
Cognitive and content engagement in game actions As discussed above, the 375
design conjecture of gameplay in E-Rebuild was that players would purposefully 376
engage in architectural building types of game actions, solve problems framed by 377
each game action using strategy and knowledge, and consequently experience a vari- 378
ety of action-based mathematical thinking processes. These thinking processes, cor- 379
responding with game actions (see Table 4.1), can involve analyzing the underlying 380
numerical, logical, and structural essentials of each problem, representing and apply- 381
ing conceptual knowledge to the problem, and then engaging in reasoning and finally 382
proof. As such, the gameplayer would develop the targeted knowledge and skills to 383
perform mathematical problem-solving or engage in mathematical thinking. Game 384
actions in E-Rebuild essentially necessitate both cognitive and content engagement. 385
Cognitive engagement in the game-based setting refers to engaging in generic cogni- 386
tive endeavors and making cognitive investment in tasks with purposiveness and 387
strategy use (Ge & Ifenthaler, 2017; Ke et al., 2016). Content engagement refers to 388
389 engagement in which players also establish positive value systems and get motivated
390 in mathematical content (i.e., concepts, relations, formulas, and theorems) process-
391 ing, representation, and application (Ke et al., 2016).
392 The user-testing findings with our initial prototypes of E-Rebuild game actions,
393 however, indicated that many players by default lacked a disposition of looking at
394 an architectural building or design problem in a mathematical way, or looking for a
395 logical explanation, which is a key prerequisite of mathematical thinking. Instead,
396 they could seek task-irrelevant gameplay, such as exploring the game world without
397 purposiveness or creating their own play experiences with the game objects (e.g.,
398 using the measurement/marking tool to sketch on the ground rather than surveying
399 the construction site and imagining the construction site as an arena and maneuver-
400 ing 3D building materials as armaments or vehicles). They might tackle game
401 actions and the associated problems in a mindless way, such as guessing with trial
402 and error. They would avoid math-relevant planning or strategic thinking, lacking
403 the acts of pre-estimating the construction cost during a trading action or predeter-
404 mining the structural properties of a construction in a building action. They would
405 also shy away from calculating the exact amount of resources available and needed
406 during an allocation action while resorting to testing and refining rough estimates,
407 randomly or with a content-generic strategy. They spent limited time reflecting on
408 or reasoning with the basis and consequence of major gameplay inputs or the con-
409 nection between them, thus lacking the practice of inductive or deductive reasoning.
410 More critically, we observed that student players generally lacked initiative in read-
411 ing during gaming, showing reluctance to peruse the mathematical problem or con-
412 tent information in a task narrative or the in-game help panel.
413 In summary, we found that cognitive and content engagement by players are not
414 promised even when the core game actions embody or operationalize the salient
415 elements of architectural and mathematical problem-solving. An earlier user-testing
416 study indicated that the percentage of cognitive engagement of players in all game-
417 play actions was 41% and that of content engagement was only 29%. To improve
418 the efficacy of E-Rebuild game actions in requiring cognitive and especially content
419 engagement, we experimented with the following design strategies for both game-
420 play actions and rules: (a) prioritizing “cognitively active” gameplay inputs and
421 behaviors that extrapolate the deed of representing or employing mathematical
422 knowledge and skills and curtail the chance of guessing and other mindless strate-
423 gies; (b) rewarding mathematical thinking (planning, calculation, and reasoning), or
424 domain-specific gameplay, as the most justified gameplay strategy by players; and
425 (c) presenting in-game mathematical information via visuals, action feedback, or
426 properties of interactive game objects, besides background narratives.
427 Prioritizing “cognitively active” gameplay inputs and conducts that extrapolate
428 the deed of “doing” mathematics and curtail the chance of guessing or other mindless
429 strategies is a major design strategy. An underlying index for our selection and refine-
430 ment of a core game action is whether and to what extent the action will n ecessitate
431 cognitively active gameplay inputs. Specifically, we have prioritized the “producing”
432 type of gameplay input (such as generating or constructing a numeric or quantitatively
accurate solution) to “choice-making” type of input (e.g., inspecting a given set of 433
options to choose a solution). We found the latter expedited and increased the usage 434
of trial-and-error strategy by players. We also tried to prioritize the game action of 435
which the associated challenge (or problem) requires the actual deed of exploiting the 436
resources of mathematical knowledge and skills, rather than just heuristics strategies 437
or general reasoning abilities, during problem-solving (Schoenfeld, 1985; Stacey, 438
2006). These design decisions are inferences extrapolated from the iterative infield 439
testing of game action archetypes, as the following example shows. 440
The learning objective of the game action of “collection,” for example, is to 441
enable the action of identifying and representing the numerical and structural prop- 442
erties of everyday objects and phenomena (e.g., shapes and measurements of con- 443
struction materials, structure of a family with the ratio of adult to child in their 444
minimum living space needs). Our original design assumption was that players 445
would proactively perform a variety of relevant learning acts (as depicted by an 446
initial design note below) enabled by and during the collection action, with the pro- 447
vision of a measurement tool, a goal statement for the collection action (e.g., using 448
containers to construct a shelter), and visual/verbal cues (e.g., a family cartoon sig- 449
nifying the composition of family member, a mathematical statement on the unit 450
space need). 451
Exploration, including measuring and appraising the landscape, is part of material gather- 452
ing. We could integrate site analysis and terrain judging (e.g., on what part of the site is 453
higher or has solid ground for building?). 454
Decision-making: What and how much materials to gather, given design needs and the 455
limited time and capability (limited storage space or limited tool resource). How can we 456
softly push a player to perform accurate calculations when gathering materials in the game? 457
Transformation: If the geometry forms or elements exist somewhere, simply gathering 458
them is a lucky act. The problem is that if one needs to use a complicated component, he or 459
she will need to comprehend and use the complex geometry or geometry transformation. 460
How can one decompose and recompose diverse geometry components? 461
The infield observation, however, indicated that most players treated the collection 462
action literally as the “treasure hunting” gameplay. They focused on searching for 463
objects dispersed in the game world, with few purposefully identifying, checking, 464
or measuring the dimensions and other structural properties of the objects collected. 465
They did not fully process the goal statement or attend to the cues on the structural 466
or numerical properties of game objects, demonstrating the intention to treat this 467
information as trivial to their “at-hand” game objective—gathering all objects to be 468
gathered. 469
These observations suggested (a) a misalignment between the action-specific 470
learning and game objectives in that the instant or by-default solution to the action- 471
specific challenge is learning-irrelevant (e.g., navigating the game world to obtain 472
objects), (b) participants’ lack of motivation to tackle a game action and challenge 473
via a mathematical strategy, and (c) participants’ lack of a planning perspective to 474
anticipate the relation between individual pieces of information or individual moves 475
to an overall plan and hence short of identifying and analyzing the mathematical 476
properties of individual objects (e.g., a shipping container or a family) to fulfill 477
478 future game actions and objectives (e.g., allocating the limited space afforded by a
479 certain number of containers to a set of families). The consequential design insight
480 is that our target middle-school students typically lack attitudes, behaviors, or skills
481 expected to perform purposeful, cognitively effortful, and mathematical gameplay.
482 Motivating participants’ purposiveness and proactivity in cognitive and content
483 engagement is important for the game action design.
484 To foster the cognitive and content engagement associated with the collection
485 action, we tried prioritizing the investigation of the numerical and structural essen-
486 tials of each item to be collected, rather than item-hunting in the virtual world, as the
487 primary action-specific challenge. For example, the player had to measure and com-
488 pute the living area of a shipping container and the living space need of a family
489 before the container and the family can be gathered. We tried adding the “produc-
490 ing” type of user input that requires the actual deed of exploiting the resources of
491 mathematical knowledge and skills, such as prompting the player to accurately com-
492 pute and inscribe the numerical values of the numerical and structural properties of
493 the collected items. We also experimented with the strategies of externalizing,
494 chunking, and presenting the sub-goals of the multistep game task, in an attempt to
495 cue the player on the integration of the at-hand moves and individual information
496 pieces to the future game actions and objectives. These design strategies have evi-
497 dently improved the players’ time spent on and the frequency of enactments of
498 mathematical information processing and reasoning processes, as evidenced by both
499 the game-logged task performance and gaming behavior analysis (Ke et al., 2017).
500 Rewarding domain-specific gameplay (i.e., mathematical planning, calculation,
501 and reasoning) as the most effective gameplay strategy for players is critical for
502 motivating game-based mathematical practice. As the literature of game-based
503 learning (Hamlen, 2012) discussed and we observed in our own design experiments,
504 the goals and values of individual uses are not necessarily allied with the designers’
505 expectations or conjectures. A hypothesis of learning games is that gameplay should
506 be both result and process driven—achieving game objectives and the processes of
507 problem-solving to achieve the objectives are both important and enjoyable. Yet
508 students frequently prioritize the result state (e.g., “passing this level”) to different
509 challenges, sometimes to an extent that they would try to bypass difficult tasks
510 instead of working through them, adopt certain “cheating” strategies, or give up
511 when a task was difficult. Fostering a positive disposition and the grit to tackle a
512 game challenge via a mathematical strategy is therefore a critical goal and condition
513 of game-based learning. In the following subsection, we describe three salient
514 themes that emerged from our design research and governing game action-specific
515 constraints and rewards that helped to motivate E-Rebuild participants’ attitudes
516 and behavior change in terms of mathematical thinking and practices.
517 From fluky psychology to mathematical reasoning During interviewing, we found
518 that participants’ pre-study gaming experiences appeared to foster a fluky psychol-
519 ogy or belief in randomness during game-based problem-solving. An exemplary
520 participant quote is, “This (E-Rebuild) is different from Minecraft. (In) Minecraft
521 you do whatever you want. This is picky.” Fluky psychology was observed in the
522 following gaming behaviors:
(In a “trading” game action) Bob1 repeated the same guessing and trial-and-error strategies 523
as before. He read the task description in less than 5 seconds and did not attend to the item 524
price bulletin inside the store. He entered random numbers (for the bid of a construction 525
item) into the ordering area, executing a trial-and-error strategy. He tried five times and 526
realized that it was not an easy guess after repetitive failures, before finally returning to the 527
virtual store to read (process) the price bulletin. 528
Bryan continued the similar trial-and-error play with the trading action. His neighbor com- 529
mented, “This is Bryan! You don’t care.” Bryan laughed and started to refine his play strat- 530
egy. In the ordering area, he placed 10 as the number of bricks to be ordered, thought for a 531
minute, and then put 45 as the total cost (which is an accurate answer, reflecting his com- 532
prehension and application of the concepts of unit price and percentage embedded in price 533
bulletin—“12% discount for ordering 6+ bricks”). He successfully obtained 10 bricks via 534
mathematical reasoning. Yet when he proceeded to purchase the next construction item, he 535
resumed guessing plus trial and error, by entering an estimated initial value and then 536
increasing it by 5 post each failed trial. He spent around 3 minutes guessing, got stuck, 537
appeared frustrated, and still showed no intention to review the price bulletin. 538
In the above examples, guessing coupled with trial and error was the primary play 539
strategy. The participants appeared to dodge mathematical reasoning and calcula- 540
tion during the trading action, until being prompted or when they realized that math- 541
ematical practices were indispensable or more efficient for problem-solving. 542
In some other cases, participants were engaged in logical reasoning without 543
applying mathematical knowledge or accurate calculation, as the following exam- 544
ples demonstrated. 545
Andy placed one of the smallest and one of the largest families into the first container, while 546
inspecting the family inventory chart during the allocation action. For the second and third 547
containers, he continued the same strategy – matching the smallest family with the largest 548
one in the to-be-assigned list and placed the pair to each container. He commented, “I think 549
I get it. So I still have four more families and three more containers.” He finished allocating 550
each of the remained largest family to each of the remained two containers to pass the level, 551
“Oh, Yay! Oh, Yay! I beat it! Oh, Yay!” He was very excited, waving and clapping the 552
hands. 553
George started by assigning the biggest family to each container. He did not process the 554
statements on the unit and ratio governing family members’ living space needs in the task 555
description. Neither did he perform any mathematical calculations. He tried to place two 556
second-largest families into another container, failed to do so, and then replaced one with a 557
smaller family. For the next container, he tried adding the second-largest family, and then 558
the smallest family. He repeated the same strategy for the fifth container. He went to the last 559
container and remarked, “I think I am going to do it this time!” He managed to add the last 560
family to the last container, and got a Congratulations screen, “YES!” His neighbor asked, 561
“How did you do it? How many (containers) did you put?” He answered, “Put the big one 562
(container) on the bottom, and the fairly big one on the top.” He did not realize that the two 563
containers he referred to actually have the same living areas. 564
As shown above, participants were involved in generic logical reasoning about 565
the space allocation, such as evaluation, information mining, planning, testing, 566
and solution refining. Yet these reasoning processes did not involve ratio- and 567
area-related mathematical calculations. 568
1
All participants’ names cited in this paper are pseudonyms.
569 Notably, there was a pattern of failure-driven learning among participants:

570 recurrent failures resulting from fluky psychology, guessing, or generic reasoning
571 would trigger the acts of mathematical knowledge application and calculation. The
572 game behavioral analysis data indicated that the percentage of explicit content
573 engagement in the participants’ gaming actions usually doubled from the earlier
574 gaming sessions to the later ones.
575 Prioritizing problem-solving efficiency There was an observed association between
576 the perceived value of mathematical solutions and participants’ engagement in
577 endorsing those solutions for game problems. In other words, when the evaluation
578 of game performance prioritized efficiency, it motivated the players’ exploration of
579 mathematical solutions. Such a pattern is illustrated by the following example.
580 Rather than estimating or calculating the amount of bricks and other building materials
581 beforehand, Sandy was ordering those items as needed during the construction process. She
582 would order four bricks initially, stacked them as a row along the foundation, and then went
583 back to order another four bricks.
584 What Sandy did was repeated by multiple other participants. Game conversations
585 among participants indicated two main reasons underlying the act of repetitively
586 ordering a small set of items: (1) avoiding the act of calculating the amount of
587 needed items based on the unit size of an item and the targeted living area of the
588 structure to be built and (2) avoiding the processing and application of percentages
589 (e.g., 6% discount for ordering 6+ bricks each time) for the cost calculation. These
590 participants apparently deemed preplanning for building (that frames mathematical
591 thinking and calculation) as effortful and unnecessary when tackling a game
592 challenge.
593 The dodging of mathematical planning and reasoning, prominently, was curbed
594 by the introduction of action-specific rules that prioritize problem-solving effi-
595 ciency—the “time limit” and “material credit” as the reward or evaluation criteria of
596 a building action.
597 Sandy managed to complete the four walls surrounding the foundation along with the door
598 and two windows. She went back to the virtual store to order more bricks to build the roof,
599 but got a pop-up window notifying that she had run out of the material credit. The second
600 time when she tried the task, she started to order 10 bricks each time and computed the cost
601 with the discount rate included during the trading action. She was almost done with build-
602 ing, yet failed the level again for exceeding the time limit. The third time she tried the task,
603 she read the price bulletin board carefully, used a paper sheet to preplan and calculate the
604 number of bricks used for each side of the wall around the foundation, pre-ordered the
605 doors and two windows, and entered an accurate value for the cost of a batch of bricks in
606 the ordering area. This time Sandy managed to pass the game level. And while she
607 complained to her neighbor how challenging the task was, she instantly went on to play the
608 next game level/task.
609 As Sandy’s case shows, only when mathematical planning, calculation, and reason-
610 ing are warranted for gameplay and perceived as high-paying strategies to tackle
611 game challenges would they be implemented. Game rules that prioritize efficiency
612 and preplanning in the action-specific evaluation would frame the associated
mathematical solutions as the most anticipated strategy and hence help to foster 613
participants’ situational interest, commitment, and potentially a positive disposition 614
toward mathematical problem-solving. 615
Reward versus punishment with trade-off between ruling and inclusiveness When 616
experimenting with game action-specific rules, we found that setting a time limit 617
with a standard threshold is challenging, because individual learners differ in both 618
play fluency (e.g., of maneuvering 3D objects or comprehending the gameplay 619
rules) and math problem-solving competency. A standard time limit tended to pun- 620
ish learners who do not have much gaming experience or are not confident about 621
game-based math problem-solving. Frequently, participants complained that they 622
were just “one brick away from the success” when they were timed out. Although 623
failure-driven learning appeared to promote cognitive and content engagement in 624
gameplay as discussed above, iterative failures due to the time limit appeared to 625
discourage participants who needed scaffolding or a lenient environment for alter- 626
ative types of game-based learning. Some participants, for example, were observed 627
working on the game problems offline using paper sheets, exchanging strategies 628
with peers, or asking for help from a facilitator or teacher amidst gameplay. These 629
actions, though potentially extending the on-task time, are part of game-based learn- 630
ing and valued cognitive or affective support for the participants. 631
A similar issue is the level of “forgiveness” toward the minor defect in the player’s 632
action-specific performance. When evaluating the players’ design solution against the 633
preset evaluation parameters (e.g., thrifty in material cost and time, reproducing the 634
shape, size, position, and functionality of the target structure and the resource usage 635
rate), setting the necessary degree of accuracy relates to the gauge of an optimal chal- 636
lenge for learners. For example, in a shelter-building task, the game evaluates whether 637
a built structure is adequately secured against a sandstorm, by testing whether any 638
single particle would blow through the walls/roof to touch the foundation. The size of 639
the particle can be up- or downscaled to test the strength of the structure. We found 640
that a low level of “forgiveness” tended to punish the participants who have a lower 641
level of fluency in maneuvering 3D objects (e.g., bricks) but comparable in the action- 642
related math competency than others. In other words, these participants could pro- 643
duce the output that is accurate in its mathematical properties but defective in its 644
design features (e.g., having a half-inch seam in between two bricks on the roof) due 645
to an inept performance of brick stacking. Executing a standard level of accuracy 646
across all design evaluation parameters would hence create a conflict between content 647
and technical competencies for individual learners. Moreover, we also found in cer- 648
tain cases, a high-fidelity evaluation of certain design parameter (e.g., the sturdiness 649
of the structure or whether it will collapse given a seismic wave) would involve the 650
evaluation and hence the exploitation of the resources of physics knowledge (e.g., 651
force, friction, elasticity, mass, and joint) that is beyond the project scope and the 652
expected knowledge base of the middle-school users. 653
Consequently, we have refined the rules on time and material cost as rewarding 654
rules rather than the pass-or-fail punishment. Specifically, badges of different criterion 655
categories (i.e., time and material credit) will be awarded to players who outperformed 656
657 the threshold value in each category. A relatively lenient, minimum requirement of
658 design evaluation is set to encourage learner agency in the initial game levels, with the
659 level of accuracy gradually scaled up in more advanced game levels when the players
660 are expected to develop an appropriate level of fluency with gameplay in addition to
661 game-based math problem-solving. The design parameters that leverage an out-of-
662 scope knowledge base were detached from the ruling base for game action evaluation.
663 The infield testing of the refined game action rules on participants’ game-based prob-
664 lem-solving and learning has supported our design decisions (Ke et al., 2017).
665 4.3.3 Designing Intermediary User Interface
666 Learning in E-Rebuild is to proactively interact with, interpret, and coordinate mul-
667 timodal elements of a building-oriented mathematical phenomenon or problem, by
668 selecting, studying, and maneuvering these elements to map out their relationship
669 and finally configure the problem solution. The interface that enables players’
670 active and meaningful interactions with the problem elements (or game objects) is
671 hence a main facet of our design exploration and has been iteratively developed,
672 tested, and refined.
673 An important user interface in E-Rebuild is the controls for viewing and navigat-
674 ing the 3D game world. To enable multiple perspectives and alternative modes of
675 inspecting game objects (or game-based math problem elements) and hence foster
676 representational flexibility in game-based math learning, we had to leverage both
677 keyboard and the mouse for the control of navigation: using arrow or WASD keys
678 for moving, right click and drag to turn/rotate the direction/angle of viewing and
679 moving, and the scroll (mouse) wheel to zoom in/out. These navigation controls, we
680 found, were novel to users who are not familiar with 3D gaming or virtual world
681 environments. Middle-school players who were used to touch pads faced a learning
682 curve in controlling the keyboard and mouse simultaneously, though they all man-
683 aged to master and use these controls naturally after playing through five to six
684 orientation levels. The comments like “change your (viewing) angle,” “turn the
685 position of your camera,” or “move up and to the other side” were frequently heard
686 during the players’ game talk, when they tried to mentor each other on how to
687 inspect or enter an object or structure in the game. These comments reflect their
688 effort and cognizance in interpreting, tracking, and positioning objects or
689 benchmarks in this 3D virtual world. Those user controls, though not very intuitive,
690 activated a conscious practicing and reasoning process of the participants in inscrib-
691 ing spatial relationships among 3D structures or objects while shifting the perspec-
692 tive and field of view. That is, they may lack technical efficiency but present
693 pedagogical or learning affordance for the learning game.
694 Such a pattern governing the integration and trade-off between technical and
695 pedagogical usability (i.e., being intuitive versus cognizant) is also observed for the
696 design of user controls for the building actions, such as moving, rotating, stacking,
697 joining, and copying/pasting building items. Different from other 3D sandbox
games, object maneuvering and building in E-Rebuild need to follow the basic laws 698
of physics to reflect the “rebuild” theme; thus the objects will collide or rebound 699
with other objects. Moving, rotating, and stacking physical objects in the xyz coor- 700
dinate axis system to constitute a structure are indeed challenging for middle-school 701
students. Participants in the initial gaming sessions were frequently found either 702
struggling in selecting the appropriate coordinate axis for moving or rotating an 703
object, stressed out when having to frequently shift the viewing angle to move and 704
position the object accurately to the target spot in a 3D space, or frustrated when 705
trying to stack the object over others yet accidently colliding objects. But there was 706
an obvious increasing trend in participants’ fluency and accuracy levels in maneu- 707
vering the 3D objects during gameplay. The gaming behavior analysis indicated that 708
the average error rate (the percentage of failed or unintended trials in all observed 709
rotation acts) of the user control with object rotation was 0.40, and the error rates 710
with other building actions (e.g., moving and stacking) were 0.15 to 0.16. The high 711
error rate with the object rotation control, on one hand, could be due to the low 712
technical usability of the interface—it needs time for the player to take on the key- 713
board shortcuts (“Tab” for shifting among the three axis coordinates, “,” and “.” for 714
clockwise or anticlockwise rotation). On the other hand, it showed that the middle- 715
school students were in need of training in the 3D rotation task for the spatial skills 716
development. We found that the intermediary interface actually had necessitated a 717
conscious performance of preplanning and extrapolating the specific axis and direc- 718
tion to rotate or move an object before actually trying them out, since instinctive 719
guessing would involve significantly more clicks and hence more effort for the 720
player. The infield studies have demonstrated that E-Rebuild participants, in com- 721
parison with the control group students, demonstrated improved mental rotation 722
task performance (Ke et al., 2017). 723
To improve the technical and pedagogical usability of the intermediary interface 724
of building, we added a joining and a copying/pasting function. These two functions 725
enable the players to join neighboring items into a panel or a structural subsystem, 726
copy/paste it to create multiple panels or subsystems, and then compose them into a 727
complicated structure. This “panel” approach, in comparison with the “brick-by- 728
brick” approach, is not only more efficient but also fostering the performance of 729
design planning before building. For example, when building a multiroom adobe 730
house or a stadium bench, the players would find it difficult to adopt a “brick-by- 731
brick” approach. When they resorted to the “panel” approach, they got to analyze 732
the structural subsystems of the house or the bench (e.g., foundation, walls, and 733
roof; support structure, seating, and railings), calculate the configuration of the sub- 734
systems (e.g., number of bricks, size, and shape of each subsystem), and gauge the 735
number of bricks needed in the inventory to copy/paste a panel or substructure. In 736
other words, the players have been engaged in composing and decomposing three- 737
dimensional shapes and building-related mathematical calculation, when using 738
joining and copying functions for building. It should also be noted that both func- 739
tions involve only left clicks (to select the neighboring items to be joined) that are 740
assisted with visual cues (with the selected items highlighted) and are found high in 741
technical usability (with a low 0.07 error rate average). 742
743 In certain cases, we had to prioritize the learning affordance of a user interface to
744 its technical usability. In the allocation action, for example, we found that the left
745 click, button control (e.g., “+” and “-” buttons) for adding or removing an item to an
746 enclosed space (e.g., allocating families to the shelter or fishes to a pond) was signifi-
747 cantly and positively associated with the guessing and trial-and-error gaming strate-
748 gies. Instead, the mechanism of drag and drop the items by unit from a personal item
749 inventory to the target space tended to reduce the behaviors of wild guessing. It is
750 because the drag-and-drop act is more dynamically linked with the visualized conse-
751 quence (e.g., a four-member family being dragged/dropped to physically occupy part
752 of the space), thus promoting participants’ awareness of the relationship between
753 occupants and the space to better comprehend the math problem to be solved.
754 4.4 Game World Design
755 In our design and research of E-Rebuild, we have focused on particularizing a back-
756 drop mission via the spatial narrative, or environmental storytelling, of the game
757 world in which the narrative element is infused into a space that a player navigates
758 through and interacts with. The game world presents multimodal, contextualized
759 representations of a mathematical problem via a series of background scenarios or
760 landscapes, interactive game objects and structures, game characters, and land-
761 marks. It helps to (a) legitimize and motivate architectural design-based mathemati-
762 cal problem-solving activities and (b) embody a multimodal problem space that
763 portrays dynamic beginning, intermediate, and goal states in consequence of the
764 player’s problem exploration and solving actions.
765 4.4.1 G
ame World as the Scenario for Contextualized
766 Problem-Solving
767 The overarching mission in E-Rebuild is to design and rebuild a disaster-damaged

768 space to fulfill the design parameters and varied residential needs specified by vari-
769 ous architectural scenarios. These architectural scenarios were presented via diverse
770 landscapes (or environment settings) along with associated “rebuild” themes, such
771 as a post-quake island with a shelter to be built, a desert with adobe houses threat-
772 ened by the sandstorm, an urban school site with portable classrooms and a stadium
773 to be restored, or a rural farm with livestock shelters to be constructed against preda-
774 tors. These environment settings, with the focal points (or landmarks) and other
775 objects (e.g., building materials, tools, and other structures and props) on the set,
776 were designed and developed by the architect and game designer collaboratively.
777 The architect in the team first designed the layout and visual content of a series of
778 environmental settings using Google SketchUp (as Figs. 4.3, 4.4, and 4.5 exempli-
779 fied). The game designers then simulated and developed these settings in Unity 3D.
4.4 Game World Design 97
Fig. 4.3 Desert and historical rural pueblo
Fig. 4.4 Rural farm
When developing environmental settings, we tried to reproduce the similar level 780
of fidelity and visual aesthetic appeal of the architect-designed sets and landscapes. 781
Yet when we infield tested the prototype with the students at the local school district, 782
the high-fidelity game worlds loaded very slowly using the school network and was 783
conflicting with the older version Chromebooks used prevalently in the school class- 784
rooms. To better design the game as a learning tool compatible with the existing IT 785
Fig. 4.5 Urban school
786 infrastructure of the school setting, we scaled down the level of fidelity as well as the
787 scope of map and adjusted the visual presentation as a conversion between the origi-
788 nal design and the typical sandbox game world (e.g., Minecraft). We found that a
789 lower level of fidelity in the game world design did not reduce the engagement level
790 of the participants. Actually, the downscaled map and game space helped to focus
791 players’ attention on the at-hand task rather than exploring the virtual world. For
792 example, a heat map analysis with the players’ navigation behaviors in the original
793 game world demonstrated that the most frequently visited or explored parts of the
794 set were not the construction site or the focus point but boundary spots (e.g., under-
795 water ditches or mountains) that were irrelevant to the problem or learning task.
796 Such an off-task game space exploration behaviors dropped by 50% in the simpli-
797 fied game world version.
798 In addition to the implicit storytelling—background scenario presentation—via
799 the game space, we used a task panel to present a description of the game environ-
800 ment story associated with a game episode (comprising multiple levels with the
801 same set/landscape) and game level-specific objectives. We tried to make the task
802 description concise, dialogic, and chunked because the infield testing indicated that
803 middle-school players tended to spend least time reading texts on screen during
804 gameplay.
805 4.4.2 Game World as the Problem Space
806 A theoretical conjecture for E-Rebuild design is that meaningful interactions with
807 multiple forms of external representations will convert an individual’s mental repre-
808 sentation for mathematical problem-solving to the sense-making process in which
References 99
relevant cognitive structures are experienced richly, multi-encoded, and integrated 809
flexibly. Given this conjecture, the game world will act as the externalized problem 810
space that animatedly portrays the beginning, intermediate, and outcome states of 811
problem-solving, via interactive game objects and dynamic visualizations (e.g., 812
visual cueing, feedback, and altered states of objects) in response to players’ inter- 813
actions with these objects. 814
In E-Rebuild, each math problem is represented in multimodal forms in the game 815
space. These multimodal external representations include (1) a verbal/syntactic task 816
description, including mathematical and design vocabulary and syntax; (2) pictorial 817
presentations, including interactive 3D game objects (e.g., varied types of construc- 818
tion items and the target structure), 2D diagrams (e.g., a floor plan), and a spatial 819
configuration of the landscape designating the location, size, and direction of the 820
structure to be built; (3) formal mathematical notations embedded in the task 821
description, inscribed onto game objects, or presented via cursor-on-target cues; 822
and (4) concrete stimuli, such as a ruler that measures distance and angles. 823
Given these multiple external representations in the game world, students must 824
actively investigate, transform, and integrate them into a coherent, internal problem 825
representation to arrive at the solution. In E-Rebuild, core game actions, such as site 826
survey, item collection, trading, building, and allocation, are designed to necessitate 827
active encoding of the external mathematical representations. The gaming objects in 828
the game space, in addition to the game actions and rules, collectively frame and 829
foster game-based mathematical problem representation and solving processes, 830
ranging from processing and selecting information needed to selecting, testing, and 831
refining mathematics-related problem-solving strategies. In summary, students’ 832
internal mathematical representations can be altered with each interaction with the 833
game world and interactive game objects and linked automatically and continuously 834
to one another to facilitate mental model development. 835
References 836
Ang, C. S. (2006). Rules, gameplay, and narratives in video games. Simulation & Gaming, 37(3), 837
306–325. 838
Garris, R., Ahlers, R., & Driskell, J. E. (2002). Games, motivation, and learning: A research and 839
practice model. Simulation & Gaming, 33(4), 441–467. 840
Frasca, Gonzalo (1999). Ludology meets narratology: similitude and differences between (video) 841
games and narrative. Retrieved from http://www.ludology.org/articles/ludology.htm 842
Ge, X., & Ifenthaler, D. (2017). Designing engaging educational games and assessing engagement 843
in game-based learning. In R. Zheng & M. K. Gardner (Eds.), Handbook of research on serious 844
games for educational applications (pp. 255–272). Hershey, PA: IGI Global. 845
Habgood, M. P. J., & Ainsworth, S. E. (2011). Motivating children to learn effectively: Exploring 846
the value of intrinsic integration in educational games. The Journal of the Learning Sciences, 847
20(2), 169–206. 848
Hamlen, K. R. (2012). Academic dishonesty and video game play: Is new media use changing 849
conceptions of cheating? Computers & Education, 59(4), 1145–1152. 850
Järvinen, A. (2008). Games without Frontiers: Theories and Methods for Game Studies and 851
Design. Tampere, Finland: Tampere University Press. 852
853 Jenkins, H. (2002). Game design as narrative architecture. In P. Harrington & N. Frup-Waldrop
854 (Eds.), First person. Cambridge, MA: MIT Press Retrieved from: http://www.anabiosispress.
855 org/VM606/1stPerson_hjenkins.pdf
856 Kafai, Y. B. (1995). Minds in play: Computer game design as a context for children's learning.
857 Routledge. AU2
859 Educational Technology Research and Development, 64(2), 219–244.
860 Ke, F., Xie, K., & Xie, Y. (2016). Game-based learning engagement: A theory-and data-driven
861 exploration. British Journal of Educational Technology, 47(6), 1183–1201.
862 Ke, F., Xu, X., Lee, S., Moon, J., Dai, Z., Pan, Y., Shute, V., Clark, K., & Erlebacher, G. (pre-
863 sented 2017, April). Math learning through game-based architectural design and building.
864 Presentation at 2017 American Educational Research Association Annual Meeting, American
865 Educational Research Association, San Antonio, TX.
866 Klopfer, E., Osterweil, S., & Salen, K. (2009). Moving learning games forward: Obstacles, oppor-
867 tunities, & openness. Boston: The Education Arcade Retrieved from http://education.mit.edu/
868 papers/MovingLearningGamesForward_EdArcade.pdf
869 Lindley, C. A., & Sennersten, C. C. (2008). Game play schemas: From player analysis to adaptive
870 game mechanics, 216784. International Journal of Computer Games Technology Retrieved
871 from: https://doi.org/10.1155/2008/216784
872 Malone, T. W., & Lepper, M. R. (1987). Making learning fun: A taxonomy of intrinsic motivations
873 for learning. Aptitude, learning, and instruction, 3(1987), 223–253.
874 Miller, C. S., Lehman, J. F., & Koedinger, K. R. (1999). Goals and learning in microworlds.
875 Cognitive Science, 23(3), 305–336.
876 Richards, J., Stebbins, L., & Moellering, K. (2013). In J. Richards, L. Stebbins, & K. Moellering
877 (Eds.), Games for a digital age: K-12 market map and investment analysis. New York: The
878 Joan Ganz Cooney Center at Sesame Workshop.
879 Salen, K., & Zimmerman, E. (2004). Rules of play: Game design fundamentals. MIT press.
880 Schell, J. (2014). The art of game design: A book of lenses. CRC Press.
881 Schoenfeld, A. H. (1985). Metacognitive and epistemological issues in mathematical understand-
882 ing. Teaching and learning mathematical problem solving: Multiple research perspectives,
883 89(4), 361–380.
884 Sicart, M. (2008). Defining game mechanics. Game Studies, 8(2), 1–14.
885 Stacey, K. (2006). What is mathematical thinking and why is it important. Progress report of the
886 APEC project: collaborative studies on innovations for teaching and learning mathematics in
887 different cultures (II)—Lesson study focusing on mathematical thinking.
888 Suits, B. H. (1978). The grasshopper: Games, life and utopia. Toronto, ON/Buffalo, NY: University
889 of Tronto Press.
890 Taylor, A. (2009). Linking architecture and education: Sustainable design for learning environ-
891 ments. UNM Press.
892 Torbeyns, J., Lehtinen, E., & Elen, J. (2015). Describing and studying domain-specific serious
893 games. Cham, Switzerland: Springer.
894 Tricot, A., & Sweller, J. (2014). Domain-specific knowledge and why teaching generic skills does
895 not work. Educational Psychology Review, 26(2), 265–283.
Author Queries

AU2 Please provide publisher location for “Kafai (1995), Schell
(2014), Taylor (2009) and Salen & Zimmerman (2004)”.
Chapter 5 1
Interweaving Task Design and In-Game 2
Measurement 3
Abstract There are two important design issues related to game-based learning 4
(GBL) in school settings: (a) the intrinsic integration of content-related tasks in 5
gameplay and (b) the real-time capture and analysis of in-game performance data. 6
In this chapter, we describe an integrative design approach that is aimed to inter- 7
weave game-based task design with in-game assessment of learning. Extending 8
other GBL projects in which the mechanism of data mining for assessment was 9
created after game development, in E-Rebuild we have designed the evidence- 10
centered, data-driven assessment during the course of game design. Design-based 11
research findings on emergent core design processes and functional conjectures on 12
the approaches of task generation and evidence accumulation are discussed, with 13
support of infield observations on the implementation feasibility and outcomes of 14
various design assumptions. 15
5.1 Introduction 16
In E-Rebuild, we examined two important design issues related to game-based 17

learning (GBL) in school settings: (a) the intrinsic integration of content-related 18
tasks in gameplay and (b) the real-time capture and analysis of in-game perfor- 19
mance data. In spite of the plethora of research on GBL, the design descriptions or 20
operational conjectures on how to develop and integrate domain-specific learning 21
tasks in gameplay are scarce (Habgood & Ainsworth, 2011; Ke & Shute, 2015). In 22
addition, learning in games has generally been assessed via external tests in a post 23
hoc manner (Shute & Ke, 2012). Adopting an educational data mining approach, 24
real-time assessment of learning based on the dynamic performance of players 25
could better capture process- and performance-oriented evidence on competency 26
development (Shute, Ke, & Wang, 2017). This has the potential to drive the design 27
of personalized or adaptive game-based learning while not being intrusive to dis- 28
tract players’ gameplay or state of flow. 29

https://doi.org/10.1007/978-3-030-04339-1_5
102 5 Interweaving Task Design and In-Game Measurement
30 In this chapter, we review an integrative design approach that intends to inter-

31 weave game-based task design with in-game assessment of learning. Different from
32 other GBL projects in which the mechanism of data mining for assessment was
33 created after game development, in E-Rebuild we designed the evidence-centered,
34 data-driven assessment during the course of game design. Design-based research
35 findings on emergent core design processes and functional conjectures on the
36 approaches of task generation and evidence accumulation are discussed, with sup-
37 port of infield observations on the implementation feasibility and outcomes of vari-
38 ous design assumptions.
39 5.2 Conceptual and Design Perspectives
40 5.2.1 Content Modeling for Game Task Design
41 Games in general can be defined as organized play that is structured by a set of

42 rules and an obstacle-tackling goal (Klopfer, Osterweil, & Salen, 2009; Schell,
43 2014). A long-held concern about using computer games for learning is that stu-
44 dents may be distracted by the play part, thus not achieving the learning goals
45 (Miller, Lehman, & Koedinger, 1999). It is argued that the extent to which the
46 content is intrinsic to the game tasks and associated actions, rules, and goals will
47 influence the game’s learning effectiveness (Richards, Stebbins, & Moellering,
48 2013). Nevertheless, empirical and theoretical research examining the design of an
49 intrinsic integration between learning and game tasks is still limited and inconclu-
50 sive (Habgood & Ainsworth, 2011).
51 A recent review on the design of intrinsic integration of domain-specific GBL
52 (Ke, 2016) argued that it is critical for a learning game designers to clarify what,
53 how, where, and when learning and content will be embedded and activated by
54 gameplay actions, rules, and game world or narrative design. Successful integration
55 of learning in games is regarded as a process rooted in an exploration of content or
56 learning elements that are fundamentally engaging within an academic discipline
57 and represent meaningful interactions between the player and the epistemic
58 frames—salient activities, understanding, and value—of the target subject matter
59 (Shaffer, 2006; Klopfer et al., 2009). These engaging learning elements should pro-
60 vide players with legitimacy in content engagement (Barab, Gresalfi, & Ingram-
61 Goble, 2010). Therefore, meaningful game task design should be driven by content
62 modeling, a process of defining the ontology of the targeted subject domain with a
63 mapping-building-block notion. During the content modeling process, componen-
64 tial concepts and skills (or competencies) along with their associated resources,
65 properties, and examples (e.g., conceptual representations or skill demonstrations),
66 as well as the relationships between these componential competencies, are specified
67 and presented via cognitive mapping.
5.2 Conceptual and Design Perspectives 103
5.2.2 Integrative Design of Task and Assessment 68
From the perspectives of construct-centered assessment (Messick, 1994) and 69

evidence-centered design for educational testing (Mislevy & Haertel, 2006; Almond, 70
Mislevy, Steinberg, Yan, & Williamson, 2015), task construction operationalizes the 71
underlying cognitive theory and the proficiency model of the targeted competencies 72
and hence is a core part of the assessment design. Task construction follows the 73
operational definition of targeted competencies and the analysis of behaviors that 74
reveal the competencies. Thus the selection or construction of relevant tasks that are 75
intended to elicit those behaviors is based on evidence, and interacting with such 76
tasks serves as both the source and evidence of knowledge and skill development. 77
The literature of learning and assessment highlights the following two salient 78
claims for the integrative design of specific tasks and the overall assessment: (a) the 79
core of authentic tasks is to simulate the representative challenges and criteria of the 80
real-world application of target competencies (Wiggins, 1990) and (b) the perfor- 81
mance of complex skills (e.g., those of mathematics and science) is moderated by 82
contexts, and hence it is necessary to construct cross-contextual tasks with varied 83
difficulty and implementation contexts to effectively measure skill development 84
(Resnick & Resnick, 1992). 85
5.2.3 I n-Game Integration of Active Learning and Authentic 86
Assessment 87
Games and simulations can be instrumental in the aforementioned task and assess- 88
ment design processes because they can (a) present a broad range of complex sce- 89
narios and tasks to model the functional context of the real world, (b) enable 90
interactions with authentic tasks and contexts, and (c) achieve a comprehensive, 91
cross-contextual assessment of the complex skill. For example, the conventional 92
approach to designing math word problems typically “attempts to strip the problem 93
context of all irrelevancies, retaining only the task information needed to engage the 94
focal knowledge and skills involved in task processing” (Messick, 1994, p. 18). Yet 95
task-based skill practice and assessment are subject to the moderation of context 96
variables. Game-based challenges, presenting controllable and customizable crite- 97
rion situations to simulate varied authentic task contexts, can afford the assessment 98
of math development that ranges between situated knowledge and abstract, decon- 99
textualized understanding (Pratt & Noss, 2002). 100
5.2.3.1 Data-Driven, Game-Based Assessment 101
Process-oriented data mining and learning analytics methods, such as Bayesian net- 102
works, social networks or structural analysis, visual or graphical analysis of event 103
paths, and sequential analysis of time series, can capture the complex and open-ended 104
105 learning trajectories in a game setting. Prior research has suggested that educational
106 data mining and learning analytics can and should be used together to exploit game-
107 based performance data to inform on students’ on-task or off-task behaviors, compe-
108 tency development related to the targeted subject matter, and hence the effectiveness
109 and design of learning games. Four recent projects (Dede, 2012; Levy, 2014; Shaffer
110 et al., 2009; & Shute & Ventura, 2013) have exemplified the potential and applicabil-
111 ity of game-based assessment via educational data mining. These projects all adopted
112 a data-intensive, evidence-based approach by collecting, measuring, analyzing, and
113 reporting dynamic data about learner performance and contexts (e.g., online log data)
114 to understand and optimize learning and the environments in which it occurs. Multiple
115 methods of data analyses (e.g., quantitative psychometric modeling, network or
116 structural analysis, and path analysis) and visualization (e.g., algorithms, models,
117 network graphs, or spatial and chronical maps) were used. On the other hand, in some
118 previous projects, data-driven assessment design tends to be a post hoc justification
119 or evaluation for the game development.
120 5.3 Design-Based Inquiry
121 Adopting the design-based research approach (Sandoval & Bell, 2004), we explored,
122 iteratively tested, and refined the core elements and design features of the game task
123 and assessment of learning through multiple mixed-method case studies. The suc-
124 cessive iterations and testing played a role similar to that of systematic variation in
125 a controlled experiment (Cobb, Confrey, Lehrer, & Schauble, 2003). Specifically,
126 the design-based investigations focused on extracting design heuristics that would
127 enable an integrative design of task and in-game learning assessment, by addressing
128 the following questions:
129 • What are the core processes that define the integrative design of game tasks and
130 assessment of learning?
131 • What are the functional conjectures on the design strategies and features of the
132 GBL tasks and assessment derived from the iterative experimental findings?
133 Extensive, longitudinal data sets were collected during the course of our design
134 experiments. These data sets involved qualitative and quantitative resources, includ-
135 ing participatory observations of the project team’s design meetings, screen- and
136 video-recording of participants’ gameplay actions/reactions, game activity logs,
137 participants’ gameplay think-aloud transcriptions, interview data, and results from
138 math knowledge tests. One hundred and twenty middle-school students participated
139 in the E-Rebuild learning program and test-played the prototypes of the GBL assess-
140 ment system that was iteratively and systematically refined based on the successive
141 design-based research findings.
142 We conducted retrospective analyses (Cobb et al., 2003) with the longitudinal,
143 design experiment data to generate situated accounts of design conjectures
144 (Schwandt, 2007; Sandoval & Bell, 2004) and to specify the functional contexts of
145 those conjectures. Specifically, we conducted thematic analysis using the design
5.4 Core Design Sectors for In-Game Learning and Assessment 105
meeting records, coded the gameplay recordings to delineate the types and proper- 146
ties of major gameplay actions and reactions via behavior analysis (Ke, 2017), and 147
then extracted the descriptive data (e.g., attempts made, levels passed, and time 148
spent) from students’ gameplay activity logs. Students’ gameplay activity logs, as 149
well as their pre- and post-gaming math knowledge test results, were also used to 150
calibrate and refine the game-based learning evidence and assessment models to 151
validate the game-based assessment mechanism. 152
5.4 C
ore Design Sectors for In-Game Learning 153
and Assessment 154
The process of interweaving game-based task and assessment development in 155

E-Rebuild includes five core design sectors: (a) designing templates of tasks based 156
on the selected gaming (i.e., architectural design) actions that necessitate, as well as 157
the scenarios (i.e., overarching design missions) that frame, the practice of the tar- 158
geted competencies; (b) setting the gameplay rule and reward system that scaffolds 159
mathematical practices; (c) developing and organizing tasks based on the compe- 160
tency, evidence, and task models; (d) drafting the game logs, including the specifi- 161
cation of in-game performance observable variables, along with the assessment 162
statistical model (e.g., a Bayesian inference network); and (e) testing, refining, and 163
validating the feasibility and effectiveness of the tasks and assessment models. The 164
ordering of these core design sectors refers to their operational sequence in our 165
game design process. However, they frequently interact with each other, thus yield- 166
ing a holistic system rather than a collection of segmented processes. 167
5.4.1 D
esigning Task Templates: Action-Oriented, Scenario- 168
Based, and Math Practice-Necessitating 169
5.4.1.1 A
ction-Oriented, Contextualized, and Multimodal Mathematical 170
Problems 171
Our design-based investigation indicated that participants’ gaming actions (i.e., the 172
core part of gameplay and main behavioral unit to be tracked in gameplay perfor- 173
mance) underlie the nature of the math content (e.g., qualitative understanding, 174
numerical calculation, and expressions and equations) and hence the targeted GBL 175
actions (e.g., identification, procedure execution, and problem-solving). Via itera- 176
tive expert review and user testing of various architectural design actions (as 177
described in Chap. 4), we have settled on a set of architectural design-based actions: 178
(a) collecting and customizing construction items, (b) site surveying and building, 179
(c) material trading, and (d) space and resource allocation. After composing these 180
learning-constructive game actions and incorporating relevant backdrop design sce- 181
narios, we then developed an assortment of game task archetypes. 182
183 In E-Rebuild, each task template was designed to depict an archetype of con-
184 textualized math story problems. These problems aim to motivate learners to
185 investigate and coordinate alternative representations of the constituent math
186 parameters, variables, and relations to solve a realistic design problem. The archi-
187 tecture-themed scenarios served to contextualize the final construction of tasks
188 and the development of task narratives. For example, a task in E-Rebuild asks the
189 player to rebuild a multiroom structure by referring to a pre-earthquake house
190 model and using a minimum number of shipping containers. The player then must
191 assign different family configurations to the shelter, with the ratio of an adult’s
192 required living space to a child’s required living space being x to y (e.g., 2:1). A
193 multimodal representation of the problem is conveyed through (a) a verbal/syntac-
194 tic task narrative; (b) pictorial presentations, including 3D game objects (e.g.,
195 shipping containers of different sizes and the house model), 2D diagrams (e.g., the
196 floor plan and a material price board in the form of a table of equivalent ratios),
197 and a spatial configuration of the landscape designating the location, size, and
198 direction of the shelter to be built; (c) formal math notations, such as ratios embed-
199 ded in the task narrative, numerals inscribed onto game objects, and operations
200 presented via gameplay feedback and the help panel; and (d) concrete stimuli, such
201 as cutting and scaling tools and a ruler.
202 Given such a multimodal math story problem, the player needs to actively inves-
203 tigate and transform math representations embedded in the game world and then
204 select and integrate the information into a coherent problem representation to arrive
205 at a solution. In E-Rebuild, core game actions necessitate active encoding and coor-
206 dination of external math representations. For instance, a single “building” game
207 action involves the acts of planning (e.g., site survey or measurement, floor or land-
208 scape planning via object positioning), composing and decomposing (prisms, such
209 as joining and stacking cuboids), covering (the space, such as painting an area),
210 surrounding (a structure, such as fencing a yard), and filling (the cavity, such as fill-
211 ing a rectangular prism with unit cubes or spheres). These actions will activate par-
212 ticipants’ interactions with the verbal, graphical, and numerical representations of a
213 geometrical figure and facilitate a coordination among these mental
214 representations.
215 5.4.1.2 Core Elements and Properties of a Task Template
216 Task archetypes can be classified based on core gameplay actions, such as collect-
217 ing, building, trading, and allocating. Certain task types, such as building, could
218 also encompass sub-tasks of planning, composing and decomposing, covering, sur-
219 rounding, and filling. Each basic E-Rebuild task template integrates a list of key
220 structural elements. These structural elements include (a) the core/driving action
221 and its enabling act(s); (b) action-appropriate goal state of the problem and the
222 criteria conditions that define the satisfactory goal state (e.g., rebuilding a pre-
223 quake house model with a minimum number of items); (c) start state of the prob-
224 lem, including unsatisfied criteria condition(s) in comparison with the goal state;
and (d) obstacles to be resolved or tackled to achieve the satisfactory goal state. 225
Obstacles will consist of action-specific constraints (e.g., limited space for an allo- 226
cation action or limited materials for a building action), sub-goal hierarchy, objects 227
to be maneuvered (e.g., actable objects), and/or those to be explored (e.g., proper- 228
ties of the structure to be built and construction materials). 229
Prior research on problem structures (e.g., Kaller, Unterrainer, Rahm, & 230
Halsband, 2004) and mathematical problem-solving (e.g., Schoenfeld, 2014) has 231
shown that the difficulty of a task archetype can be predicted by the requisite math 232
knowledge and skills needed to solve the task, as well as the structural features of 233
the task, like the number of constraints—where more constraints mean that more 234
variables need to be considered during problem solution. Another task feature which 235
can influence its difficulty concerns the ambiguity of goal hierarchy (i.e., the obscu- 236
rity of goal priorities in the assembly of sub-goal states). Finally, the existence of 237
suboptimal, alternative solutions in a given task renders it more difficult to solve. 238
Overall, the prerequisite math knowledge in a given task archetype reflects the 239
required level of competency in the content domain, while the task’s structural fea- 240
tures affect math problem-solving performance, including problem interpretation, 241
problem-solving strategy selection (or planning), execution, and monitoring. 242
5.4.1.3 Composing Tasks for Game Level Development 243
E-Rebuild is a multi-episode game. Each game episode then consists of multiple 244
levels. A game level can be developed as either an architectural design problem with 245
multiple instantiations of task archetypes or as a sector of an overarching inquiry 246
(i.e., episode) which carries only one instantiation of a task archetype. We have 247
experimented with these two options of task composition for game level develop- 248
ment and found that the former was challenging for novice student players who 249
lacked skills related to problem interpretation, sub-goal generation, and analysis. In 250
addition, a typical duration of a class session (and hence the in-class gaming ses- 251
sion) was only 45 min, which made it difficult for learners to fully practice and 252
complete a comprehensive inquiry-based game level. To bolster engagement in the 253
game and learning, we found it necessary to simplify the task composition by fur- 254
ther decomposing a task within a game level. 255
Our initial game level prototype (referred to as “row house”) illustrates our 256
design decisions governing task composition. Figure 5.1 depicts the game level 257
which serves as the backdrop (or design inquiry) for the player who has to build a 258
shelter for survivors using shipping containers. The design inquiry consists of three 259
types of tasks—collecting (survivors and building materials), building (a shelter), 260
and allocating (survivors to the shelter built). During collection, the player is 261
expected to calculate the space needed to accommodate each type of surviving fam- 262
ily (i.e., the various adult-child ratios in the family composition and associated 263
space requirements), measure the living space (area) afforded by each container, 264
and then figure out the number of containers needed before collecting both survi- 265
vors and containers within a limited time frame. During building, the player is 266
Fig. 5.1 Composition of tasks in a game level
267 expected to review the location, three-dimensional shape, and size of the pre-quake
268 structure and replicate it using collected shipping containers. During allocation, the
269 player then has to allocate the various families into the limited number of container
270 rooms in the shelter. For each type of task, multiple constraints are applied, such as
271 the variety and number of types of families and containers to collect, the time limit,
272 the design criteria, and the space limits per room compared to the space needed by
273 the different families. Additional constraints may be added to further increase dif-
274 ficulty, such as minimizing the transportation cost of collecting a container.
275 In the earliest version of “row house,” all component tasks were presented as a
276 holistic inquiry without in-game feedback or cueing on the composition of sub-
277 tasks. During its infield testing, we observed that middle-school students typically
278 lacked the awareness or skill to identify and plan sub-goals for the inquiry. They
279 were quick in figuring out how to collect families and containers yet failed to pur-
280 posefully check and calculate the space needed per family or use the family collec-
281 tion results to plan the container collection act. Hence they would randomly collect
282 containers and then either not replicate the row house or build the structure but not
283 allocate all of the families given the wrong number and/or type of containers col-
284 lected. To scaffold their sub-goal generation, we decomposed the broader inquiry
into segments and had the game check the player’s collection performance before 285
allowing him to proceed to the next step/task (as Fig. 5.1 illustrates). A consequence 286
of such a linear segmentation, however, was that certain players adopted the generic 287
strategy of guessing, along with trial-and-error actions to pass the collection sector, 288
rather than processing and applying the task-related math information. Segmentation 289
also caused players to inadequately develop an overarching interpretation of the 290
problem. For instance, they failed to recall the living space needs of the different 291
families (i.e., information identified in the collection sector) during the allocation 292
sector. We additionally experimented with task-equivalent game level development, 293
in which each component task represented a separate level. The component tasks 294
comprised a partial problem to solve. 295
Now, in the game, after the player completes a sequence of related task levels, he 296
then must tackle a complete problem (boss level) which is an all-inclusive instantia- 297
tion of the sum of component tasks. Correspondingly, we created two levels of game 298
chunking—level (task-equivalent) as the child or lower-level units and episode 299
(multitask-equivalent) as the parent or higher-level units. Such a mixture of task 300
composition and decomposition to particular actions needed, as the field test data 301
indicated, was well received by student players and associated with an increased 302
level of task engagement and math problem-solving performance. 303
Chunked versus holistic task We experimented and conducted a qualitative com- 304
parative analysis with the users’ participation behaviors in between the holistic task 305
with which a player had to explore and chunk the componential problem-solving 306
phases by himself and the chunked task series in which each problem was purpose- 307
fully chunked into multiple phases (or levels) that players will proceed sequentially 308
with. When interacting with the sequenced phases (e.g., material collection/trading, 309
structure building, and space allocation) in a design task, participants followed a 310
framed problem-solving path to solve and connect each part of the puzzle. The com- 311
parative analysis of participants’ game-based problem-solving processes and perfor- 312
mance in the two task structures indicated that an explicitly chunked task structure, 313
by offering partial representation of a complex problem, fostered task engagement. 314
Participants with the holistic task structure demonstrated obviously more off-task 315
behaviors than those with the chunked task structure. For example, participants of 316
the former were frequently found wandering around the game world, casually play- 317
ing with a game object (e.g., using the measurement tool to draw line sketches 318
instead of site surveying), and reporting feeling stuck or lost during gaming. 319
On the other hand, an explicitly chunked task structure presented less opportuni- 320
ties for participants to perform failure-driven, reflective learning. For example, with 321
a holistic task, participants were frequently observed recalculating the size of each 322
container, the total living space needed by the families, and hence the number of 323
containers needed, when they failed to allocate all families. Yet with the chunked 324
task structure, such a failure-driven mathematical practice occurred less frequently 325
because at the end of the building phase, the game would evaluate the shelter built 326
and alert the player if containers used were not enough. In other words, the game 327
had done a critical part of the problem representation and solution for the player. 328
329 Connection between tasks For multilevel episodes in E-Rebuild, we explored

330 alternative ways of constructing and relating game tasks or levels. For example, in
331 an “Island” game episode, every level presented a task following the same problem
332 template, while the specifics of core problem parameters (e.g., the size of a con-
333 tainer, the number and type of families, the adult-child ratio in space needed) varied
334 to create incremental difficulty in mathematical problem-solving. In a “Desert” epi-
335 sode, a sequence of levels represented a collection of tasks governed by an over-
336 arching architectural design scene. Across the levels there was an incremental
337 inclusion of new parameters (e.g., new building and site design criteria) in the prob-
338 lem to create incremental complexity in problem-solving. The gaming behavior
339 analysis and infield observation did not reveal any obvious difference between the
340 two task or level arrangements in relation to players’ task engagement or game-
341 based problem-solving performance.
342 Instead, we observed that the degree of granularity required by a building task
343 (e.g., building a structure using prefabricated shipping containers versus using 2-by-
344 2-meter blocks) in different game episodes moderated players’ performance in 3D
345 object maneuvering. It was reported and observed that the acts of rotating and stack-
346 ing 3D objects in a higher granularity level were more time-consuming than that in
347 a lower granularity level. The challenge of manipulating 3D objects during building,
348 interestingly, appeared to foster the behaviors of preplanning for building, such as
349 predicting the rotation and location of a block before placing and stacking it onto
350 another and gauging the minimum space needed or the most efficient way of build-
351 ing. These behaviors were associated with more observed events of mathematical
352 reasoning and calculation.
353 5.4.2 S
etting Gameplay Rules to Motivate Content-Related
354 Task Performance
355 Prior analyses of the video- and screen-recorded gaming behaviors of the partici-
356 pants suggested that intentionality and mindfulness were two prerequisite facets of
357 game-based engagement. Intentionality refers to the ability to create and maintain
358 goal-directed gaming behaviors. Mindfulness refers to the level of reflective think-
359 ing (e.g., diagnosing the reason of a failed game action) and degree of planning
360 (e.g., comprehending the purpose of a future game action) during gameplay. The
361 analyses with the gaming behaviors and game logs indicated a positive association
362 between the presence of intentionality and mindfulness in participants’ gaming
363 behaviors and game level completion (Ke et al., 2017). The two task engagement AU1
364 patterns appeared to mediate the players’ processing and application of in-game
365 mathematical information. Yet even when primary game actions and tasks were pur-
366 posefully designed to stimulate mathematical problem-solving, players did not nec-
367 essarily demonstrate intentionality or mindfulness in their gaming moves. Guessing
368 and trial-and-error behaviors were two strategies prioritized by novice players for
369 game-based problem-solving. More critically, their use of trial-and-error and
guessing lacked goal direction, mindful planning, reflection, or mathematical think- 370
ing. It appeared that novice players, especially those who were motivated to play 371
rather than to learn, were reluctant to accept the circumstance that gameplay in 372
E-Rebuild is an intentional and reasoning process; they tended to avoid much cogni- 373
tive and mathematical engagement. Therefore, a critical challenge of task design in 374
E-Rebuild as a learning game was to motivate players’ intentionality, mindfulness 375
(or cognitive engagement), and hence processing and application of task-related 376
mathematical knowledge (or content engagement) during game-based problem- 377
solving. A theoretical perspective that helps to shed the light on this design chal- 378
lenge is what Schoenfeld (2014) described about the three challenges faced by 379
students in mathematical problem-solving: 380
In some cases, “much of the mathematical knowledge that the students had at their disposal, 381
and that they should have been able to use, went unused in problem solving. This was not 382
because they had forgotten it (a matter of resources) or because they ran out of time to use it 383
(a matter of control), but because they did not perceive their mathematical knowledge as being 384
useful to them, and consequently did not call upon it (a matter of belief systems).” (p. 13) 385
Students’ problem-solving performance is not simply the product of what the students 386
know; it is also a function of their perceptions of that knowledge, derived from their experi- 387
ences with mathematics. That is, their beliefs about mathematics – consciously held or 388
not – establish the psychological context within which they do mathematics. (p. 14) 389
In general, Schoenfeld (2014) argued that mathematical problem-solving involves 390

(a) problem-solvers’ operational understanding of the mathematical information (or 391
resources at disposal); (b) techniques (heuristics) problem-solvers have or lack for 392
making progress on unfamiliar problems; (c) the way (control) they use, or fail to 393
use, the information at their disposal; and (d) their mathematical world view (belief 394
system) that determines the ways that the knowledge in the first three categories is 395
used. The lack of positive mathematical belief systems actually underlies the obser- 396
vation in our study that players were not willing or had the awareness to access their 397
mathematical knowledge to solve the game tasks. 398
Based on the infield observations and the aforementioned importance of players’ 399
positive perceptions of mathematical information during problem-solving, we spec- 400
ulated that setting gameplay rules that highlight mathematical knowledge applica- 401
tion as the most efficient, if not the only, strategy for problem-solving should 402
reinforce mindful and content-related performance in the game. Hence we designed 403
a set of task constraints and reward rules that aim to curtail absent-minded, casual 404
gaming actions and enhance players’ awareness of the connections between each 405
math skill-based game action and the reward of gaming progress or credit. For 406
example, during the aforementioned allocation task, players intermittently added 407
families to random containers and used a trial-and-error approach to reshuffle them 408
(families to containers) until the task was solved. Consequently, a material and time 409
credit deduction/penalty was then added for each allocation-reshuffling move. Once 410
the deducted credits hit some threshold, the player has to restart the task/level. With 411
this supplementary gameplay rule for the allocation task, the observed guessing and 412
trial-and-error gaming acts were reduced. 413
414 Similarly, in a trading task, the player is supposed to order building items in bulk
415 to gain a discount (in an increasing percentage depending on the amount ordered).
416 Yet players frequently found ways to circumvent the percentage calculation and
417 only order items in part, not bulk. They also did not preplan the number of items to
418 be traded before a building action; instead they would buy items as needed during
419 building. To motivate mathematical information processing and calculation, we
420 enforced a “transaction fee” for each trade made, reduced the default material credit,
421 and increased the distance (and hence transportation time) between the store and
422 construction site. We also added a badge system (in addition to the game level pass/
423 fail) that explicitly reviews and rewards the degree of accuracy and efficiency in a
424 player’s game-based problem-solving actions. All of these gameplay rules, as the
425 design-based research findings indicated (Ke et al., 2017), have subsequently moti-
426 vated and increased the frequency of observed game-based cognitive and content
427 engagement acts, reduced the number of failed game level attempts, and promoted
428 game-based task performance.
429 5.4.3 G
enerating and Organizing Tasks with the Assessment
430 Models
431 A retrospective analysis of the observational data and design artifacts from the
432 E-Rebuild design meetings suggested that game task template generation and task
433 instantiation are confined and substantiated by the development of competency, evi-
434 dence, and task models for in-game assessment of learning. The sequential record
435 of design events indicated a concurrent, interactive association among the afore-
436 mentioned task and assessment design processes.
437 In E-Rebuild, we adopted the evidence-centered assessment design approach
438 (Mislevy, Almond, & Lukas, 2003) to construct game-based assessment.
439 Specifically, we started by defining the claims to be made about participants’
440 math competencies (i.e., competency modeling), establishing what actions or ele-
441 ments of gameplay constitute valid evidence of the claim (i.e., evidence model-
442 ing), and determining the nature and form of game tasks that will elicit that
443 evidence (task modeling) (Shute, 2011; Shute & Ke, 2012; Shute et al., 2017).
444 Although assessment design flows from competency to task modeling, in practice
445 it is more iterative. Diagnosis flows in the opposite direction. That is, the learners’
446 performance (recorded by game logs) during a game level/task will provide the
447 evidence or data (e.g., logged scores of observable variables) that are passed on to
448 the competency model, which in turn updates the claims (e.g., probabilities) about
449 relevant competencies. Based on the competency claims made, E-Rebuild, in the
450 long term, can dynamically present adaptive help or personalized game tasks/
451 levels to the player.
452 The competency model, as a framework for defining the targeted skills and
453 knowledge requirement in the game-based learning system, is a collection (or hier-
454 archy) of competencies that jointly define the ultimate achievement, successful
performance, or proficiency. Mapping the task templates with the competency 455
model helps to ensure game-based learning and assessment of the targeted compe- 456
tencies and skills (see Fig. 5.2). The evidence model, presented as a Q-matrix 457
(Fig. 5.3), specifies which elements (or observed performance measures) of major 458
game task archetypes will demonstrate the practice and hence elicit the evidence of 459
competency facet(s) and how they will be combined to afford the learning and 460
assessment of the targeted competencies. The task model, presented as another 461
Q-matrix (Fig. 5.4), states which specific game level/task will provide the evidence 462
for which competency facet(s). The evidence and task models help the design team 463
to (a) estimate whether, when, and what tasks generated will accumulate enough 464
evidence and (b) gauge the difficulty, discrimination quality, and hence sequencing 465
of generated/instantiated tasks across game levels and episodes. 466
5.4.4 D
esigning Game Log with Observables to Record Game- 467
Based Learning Evidence 468
During the course of E-Rebuild design, we designed the game logs to track game- 469
based performance measures intended to provide evidence/data to propagate and 470
validate the statistical model (e.g., Bayesian network) for game-based assessment. 471
Game logs are XML files that are created at the end of gameplay. Since the goal is 472
to assess a participant’s competency from how the participant plays the game, infor- 473
mation that capture evidence of such competencies are logged in the XML file. The 474
following is the content of an example XML file for user “abc” for the game level 475
“SchoolPlacement02”: 476
Fig. 5.2 An exemplary design document depicting competency model and game-based task tem-
plate design (Shute et al., 2017)
Reason with ratio and proportional reasoning
114
Compare ratios
with whole
Task Name ObsName number Recognize a ratio Recognize a ratio Recognize a ratio
measurement relationship relationship between relationship between Represent a ratio Represent a ratio Represent a ratio Calculate the unit Recognize a percent of
using tables of between 2 quantities 2 quantities in 2 quantities in relationship via relationship via relationship via rate (a/b) associated a quantity as rate per
equivalent ratios in numerical form verbal form symbolic form numerical form verbal form symbolic form with a ratio (a : b) 100
timeToCompletion 0 1 1 1 1 0 1 1 0
Material Credit 0 0 0 0 0 0 0 1 0
Allocation Task scratchpad

editing(math related) 0 0 0 0 1 0 0 1 0
assignment
operation 0 0 0 1 0 0 1 1 0
# of trades 1 1 1 0 1 0 0 1 1
5
scratchpad
Trading Task percentage lost in
trade avg 1 1 1 0 1 0 0 1 1
cut (for resourcing) 0 0 0 0 0 0 0 0 0
scale (for resourcing) 0 0 0 0 0 0 0 0 0

structure size 0 0 1 0 0 0 1 0 1
structure location 0 0 0 0 0 0 0 0 0
structure direction 0 0 0 0 0 0 0 0 0
Building Task # copy/paste failed 0 0 0 0 0 0 0 0 0
scratchpad
ruler record 0 0 1 0 0 0 1 0 0
timeToCompletion 1 1 1 1 1 0 1 1 1
Game Task Material Credit 1 1 1 1 1 0 1 1 1
Happiness Credit 0 1 1 1 1 0 1 1 1
Fig. 5.3 Part of an evidence model (Q-matrix) example (Shute et al., 2017). Facets of the focus competency are listed in columns, and the indicators are listed
in rows
Interweaving Task Design and In-Game Measurement
5.4
Comprehend a (De)compose Find surface Compute the

ratio Solve problems quadrilaterals Compute the areas of 3D volume of right Compute area
relationship via involving finding Calculate the and polygons area and figures using rectangular and
numerical, the whole, given unit rate (a/b) into (right) perimeter of nets of prisms (V = l w circumference of
verbal, and a part or a associated with triangles and triangle and rectangles and h, V = b h) a circle using
Episode Level symbolic forms percent a ratio (a : b) rectangles rectangle (right) triangles formulas
Island Collect Container 1 0 0 0 0 1 0 0 0
Island Collect Family 1 1 0 1 0 0 0 0 0
Island Build Training 0 0 0 0 0 0 0 0
Island Fill Training 1 0 0 0 1 0 0 0
Island Build 1 1 0 1 1 1 0 0 0
Island Fill 1 1 0 1 0 1 0 0 0
Island Build 2 1 0 1 1 1 0 0 0
Island Fill 2 1 0 1 0 1 0 0 0
Island Build 3 1 0 1 1 1 0 0 0
Island Fill 3 1 0 1 0 1 0 0 0
Desert Copy Training 1 0 1 1 1 1 0 0 0
Desert Copy Training 2 0 1 1 1 1 0 0 0
Desert Placement Training 1 0 0 0 0 0 0 0 0
Desert Build 1 1 1 1 1 0 1 0 0
Desert Angle Build 1 1 1 1 1 0 1 0 0
Desert Location Build 1 1 1 1 1 0 1 0 0
Core Design Sectors for In-Game Learning and Assessment
Desert Fill 1 1 0 1 0 1 0 0 0
School Place 1 1 1 1 0 1 0 0 0
School Place 2 0 0 0 0 0 0 0 1
School Fill1 1 0 1 0 1 0 0 0
School Paint 1 0 1 1 0 0 0 0 1
School Stadium 1 0 1 1 1 0 0 0 0
School Stadium 2 0 1 1 1 0 0 0 0
School Paint 2 0 1 1 1 0 0 0 0
Farm Angle 1 0 1 1 0 0 0 0 0
Farm Perimeter 1 0 1 1 0 1 0 0 0
Farm Area 1 0 1 1 0 1 0 0 0
Farm Volume 1 0 1 1 0 0 0 1 0
16 14 25 11 17 3 1 2
115
Fig. 5.4 Part of a task model (Q-matrix) example

477 <?xml version="1.0" encoding="utf-8"?>

478 <root>
479 <Name>abc</Name>
480 <Level>SchoolPlacement02</Level>
481 <Time>40.7275429</Time>
482 <NumBlocks>0</NumBlocks>
483 <NumTrades>0</NumTrades>
484 <TotalLost>0</TotalLost>
485 <MaterialCredits>10000</MaterialCredits>
486 <LevelComplete>true</LevelComplete>
487 </root>
488 The XML file has elements with tags like root, Name, Level, and Time which are
489 enclosed within “<” and “>.” For example, <MaterialCredits>10,000</
490 MaterialCredits> is a MaterialCredits element with content 1000. The elements
491 other than root, Name, and Level are the observables for the game level
492 SchoolPlacement02.
493 E-Rebuild prototype has 34 game levels and each level has its own set of observ-
494 ables. Table 5.1 lists a few game levels along with their observables. The last entry
495 of Table 5.1 is the union of the observables from all the game levels. As it is evident
496 from Table 5.1, each game level logs only a subset of the total observables.
497 5.4.5 Testing and Refining Task and Assessment Development
498 We have iteratively tested and refined the task templates, including game actions
499 and their interfaces, along with the instantiated tasks via iterative design experi-
500 ments. To examine the affordance of game tasks in fostering math problem-solving
501 skills and explore the nature of problem-solving processes, we collected longitudi-
502 nal, gameplay data of participants in every gaming session. Data were collected via
t1.1 Table 5.1 Sets of observables for the given game levels
t1.2 Exemplary level Observables
t1.3 21ContainerCollect Time, NumWrong, MaterialCredits
t1.4 26FamilyPlacement Time, NumWrong, AssignmentComplete
t1.5 MaterialCredits
t1.6 SchoolAssignment01 Time, AssignmentComplete, NumAssignments, NumFailedAssignments,
t1.7 Num-FamilyCollected LevelComplete
t1.8 IslandBuild02 Time, NumBlocks, NumTrades, Total-lost, MaterialCredits, distance,
t1.9 size, angle BuildingComplete, LevelComplete
t1.10 All levels Angle, AssignmentComplete, building-
t1.11 Complete, distance, LevelComplete, MaterialCredits, NumAssignments,
t1.12 NumBlocks, NumFailedAssignments NumFamilyCollected, NumTrades,
t1.13 NumWrong, size, time, TotalLost
infield observations of students’ gameplay, informal interviews, game logs, as well 503
as video- and screen-captured gameplay behaviors and think-aloud protocols. We 504
conducted a qualitative thematic analysis with observation notes and gameplay 505
recordings to extract salient themes on participants’ gaming actions and their game- 506
based problem-solving processes. These qualitative themes defined the nature of 507
commonly observed and unique player actions, along with the critical properties of 508
each action (e.g., objects engaged, purpose, and relevance to mathematical problem- 509
solving). We then classified them into categorical themes depicting participants’ 510
cognitive and affective engagement states and their major gaming actions. 511
Using the categorical themes and BORIS (an event logging software), we per- 512
formed pattern matching and systematically coded the recordings of a representa- 513
tive sample of participants’ gameplay actions and reactions, using 30 second time 514
intervals as the primary coding unit. The qualitative thematic and systematic behav- 515
ioral coding results informed about the nature of game-based mathematical 516
problem-solving processes in the context of variant types of game actions and tasks 517
and illustrated the relative effectiveness of each type of game actions, tasks, and 518
gaming interfaces. Based on the analyses results, specific game actions and tasks 519
were then refined and tested iteratively across design experiments. 520
Training and Testing of the Game-Based Learning Assessment Model Using 521
Bayesian Network We have trained and tested the game-based math learning 522
assessment using Bayesian network in Netica. Here are some reasons why we 523
choose Bayesian network for in-game learning assessment in this project: 524
• It is flexible in allowing domain experts to encode the domain knowledge by 525

defining nodes (i.e., competency and observable variables) and edges (i.e., the 526
relationships among the variables) in a directed acyclic graph. 527
• It handles missing data. This feature is a requirement in our case because each 528
game level provides only a subset of the observables. 529
Specifically, we trained, refined, and substantiated the Bayesian network using 530
both users’ gameplay data and expert review (Ke & Shute, 2015). Users’ game- 531
play data, as well as external pre- and post-gaming math knowledge test results, 532
were collected during iterative design experiments in E-Rebuild. We consolidated 533
the data stored in multiple XML game logs along with the corresponding pretest 534
scores into a single CSV (comma-separated values) format file. Each entry or row 535
in the output CSV file corresponds to a particular user and her game log for a par- 536
ticular game level played at a certain time. The consolidated data file contains 537
values of different types: integers, strings (categories), and floats. Since we have 538
selected Netica’s Bayesian network with discrete nodes for the game-based assess- 539
ment of learning, all raw values were converted to categorical values. The category 540
thresholds were determined from the training data and reviewed/refined by the 541
educational measurement, math education, and game design experts in the project 542
AU2 team (Fig. 5.5). 543
Fig. 5.5 A snapshot of a Bayesian network model
544 We used Netica to create the Bayesian network described above and conducted
545 experiments on the preprocessed categorical data set. The goal was to learn a model
546 based on the training set about the relationships between observables across all
547 gameplay and the users’ competencies. Later, the learned model can take as input a
548 set of observables from a gameplay of an unknown user and predict competencies
549 of the user.
550 We fed the training data set into the Bayesian network via the Netica case file
551 format. We chose the expectation-maximization algorithm to train the model. To
552 establish initial values for the conditional probability tables in our Bayesian net-
553 work model, we considered each level played as an individual case. The initial val-
554 ues for the student’s mathematic ability are taken from their external math test
555 results. The network is trained using these values along with the log data. The net-
556 work is then tested against the posttest data.
557 After the training was completed, we continued with the testing process. We first
558 created a Netica control file which controls the output we want from Netica. We
559 then fed the testing data set from Netica to populate nodes in the Bayesian networks
560 to obtain the predictions for the participant’s mathematical competencies. An asso-
561 ciation analysis between the in-game assessment results (i.e., Bayesian network
562 predictions for individual learners) and the external post-gaming math test results
563 was conducted to validate the Bayesian network model. The correlation analysis
564 was conducted to examine the consistency between the predicted result of the cur-
565 rent Bayesian network model (i.e., low, medium, or high in the targeted compe-
566 tency) and the categorized posttest performance of middle-school gaming
567 participants. The analysis result indicated a significant association, r = 0.40,
568 p = 0.02. Error rate was used as another matric to evaluate the trained Bayesian
569 network model. The error rate measures the overall accuracy of predictions. The
570 error rate was 0.443 when we validate the model trained on Pa with Pb. This shows
571 that the model trained on a certain population can be used to make predictions on an
572 entirely different population.
5.5 Summary and Future Investigation 119
5.5 Summary and Future Investigation 573
By examining and discussing the task design and assessment modeling processes of 574
E-Rebuild, we were able to explore how the two core game design elements func- 575
tion together to develop a game-based learning system. Such a system can poten- 576
tially serve as both a learning and an assessment tool. The five core design sectors 577
illustrate the design heuristics of the learning-play integration during game task and 578
level development; data mining-based assessment models for the problem-oriented, 579
task-centered learning; and the non-interruptive assessment mechanism in a playful 580
learning environment. They also support the innate association among content/ 581
domain modeling, task template design and instantiation, and assessment model 582
development. 583
Multiple design issues remain unaddressed and warrant more design-based 584
investigation. The main issues are: 585
1. The existing game tasks are developed by the project team and cannot afford the 586
activation and assessment of varied sets of math competencies; salient parame- 587
ters of learning-centered game tasks need to be extracted to inform the genera- 588
tive method to enable cost-effective, scalable, and participatory game task 589
development by teachers and other direct users. 590
2. The mechanism that enables the real-time collection and analysis of the logged 591
gameplay data needs further development, which will involve the design of an 592
application that dynamically extracts the data from the game logs into the 593
Bayesian network or other applicable statistical models. 594
3. Incoherency still exists in designing game as a learning tool versus as an assess- 595
ment tool, in that the sequencing of tasks for learning focuses on scaffolding via 596
a content and difficulty progression and iterative practices for deep learning, 597
whereas the arrangement of tasks for assessment prioritizes selecting tasks that 598
discriminate most and presenting them to collect and accumulate evidence. 599
AU3 References 600
Almond, R. G., Mislevy, R. J., Steinberg, L., Yan, D., & Williamson, D. (2015). Bayesian networks 601
in educational assessment. New York, NY: Springer. 602
Barab, S. A., Gresalfi, M., & Ingram-Goble, A. (2010). Transformational play: Using games to 603
position person, content, and context. Educational Researcher, 39(7), 525–536. 604
Cobb, P., Confrey, J., Lehrer, R., & Schauble, L. (2003). Design experiments in educational 605
research. Educational Researcher, 32(1), 9–13. 606
Dede, C. (2012, May). Interweaving assessments into immersive authentic simulations: Design 607
strategies for diagnostic and instructional insights. In Invitational research symposium on tech- 608
nology enhanced assessments. 609
Habgood, M. P. J., & Ainsworth, S. E. (2011). Motivating children to learn effectively: Exploring 610
the value of intrinsic integration in educational games. The Journal of the Learning Sciences, 611
20(2), 169–206. 612
613 Kaller, C. P., Unterrainer, J. M., Rahm, B., & Halsband, U. (2004). The impact of problem struc-
614 ture on planning: Insights from the tower of London task. Cognitive Brain Research, 20(3),
615 462–472.
617 Educational Technology Research and Development, 64(2), 219–244.
618 Ke, F., & Shute, V. J. (2015). Design of game-based stealth assessment and learning support. In
619 C. Loh, Y. Sheng, & D. Ifenthaler (Eds.), Serious games analytics (pp. 301–318). New York,
620 NY: Springer.
621 Klopfer, E., Osterweil, S., & Salen, K. (2009). Moving learning games forward: Obstacles, oppor-
622 tunities, & openness. Boston: The Education Arcade. Retrieved 28 April 2011 from http://
623 education.mit.edu/papers/MovingLearningGamesForward_EdArcade.pdf.
624 Levy, R. (2014). Dynamic Bayesian network modeling of game based diagnostic assessments.
625 CRESST Report 837. National Center for Research on Evaluation, Standards, and Student
626 Testing (CRESST).
627 Messick, S. (1994). The interplay of evidence and consequences in the validation of performance
628 assessments. Educational researcher, 23(2), 13–23.
629 Miller, C. S., Lehman, J. F., & Koedinger, K. R. (1999). Goals and learning in microworlds.
630 Cognitive Science, 23(3), 305–336.
631 Mislevy, R. J., Almond, R. G., & Lukas, J. F. (2003). A brief introduction to evidence-centered
632 design. ETS Research Report Series, 2003(1), i–29.
633 Mislevy, R. J., & Haertel, G. D. (2006). Implications of evidence-centered Design for Educational
634 Testing. Educational Measurement: Issues and Practice, 25(4), 6–20.
635 Resnick, L. B., & Resnick, D. P. (1992). Assessing the thinking curriculum: New tools for educa-
636 tional reform. In Changing assessments (pp. 37–75). Dordrecht: Springer Netherlands.
637 Richards, J., Stebbins, L., & Moellering, K. (2013). Games for a digital age: K-12 market map and
638 investment analysis. New York, NY: Joan Ganz Cooney Center at Sesame Workshop.
639 Sandoval, W. A., & Bell, P. (2004). Design-based research methods for studying learning in con-
640 text: Introduction. Educational Psychologist, 39(4), 199–201.
641 Schell, J. (2014). The art of game design: A book of lenses. Boca Raton: CRC Press.
642 Schoenfeld, A. H. (2014). Mathematical problem solving. Orlando, FL: Academic Press, Inc.
643 Shaffer, D. W. (2006). Epistemic frames for epistemic games. Computers & Education, 46(3),
644 223–234.
645 Shaffer, D. W., Hatfield, D., Svarovsky, G. N., Nash, P., Nulty, A., Bagley, E., et al. (2009).
646 Epistemic network analysis: A prototype for 21st-century assessment of learning. International
647 Journal of Learning and Media, 1(2), 33–53.
648 Shute, V. J., & Ventura, M. (2013). Measuring and supporting learning in games: Stealth assess-
649 ment. Cambridge, MA: The MIT Press.
650 Shute, V. J., & Ke, F. (2012). Games, learning, and assessment. In D. Ifenthaler, D. Eseryel, &
651 X. Ge (Eds.), Assessment in game-based learning (pp. 43–58). New York, NY: Springer.
652 Shute, V., Ke, F., & Wang, L. (2017). Assessment and adaptation in games. In P. Wouters & H. van
653 Oostendorp (Eds.), Instructional techniques to facilitate learning and motivation of serious
654 games (pp. 59–78). Chem, Switzerland: Springer.
655 Shute, V. J., Wang, L., Greiff, S., Zhao, W., & Moore, G. (2016). Measuring problem solving
656 skills via stealth assessment in an engaging video game. Computers in Human Behavior, 63,
657 106–117.
658 Wiggins, G. (1990). The case for authentic assessment. Washington, DC: ERIC Digest.
Author Queries

AU1 References “Pratt & Noss, 2002, Schwandt, 2007, Ke, 2017,
Ke et al., 2017, Shute, 2011” are given in the text but not
provided in the reference list. Please provide.
AU2 Please confirm the inserted figure citation for Fig. 5.5.
AU3 References “Shute et al. (2017), Shute et al. (2016)” were not
cited anywhere in the text. Please provide in text citation or
delete the reference from the reference list.
Chapter 6 1
Designing Dynamic Support for Game- 2
Based Learning 3
Abstract The role of support for game-based learning cannot be overemphasized. 4

It remains inconclusive as to what, when, and how support for learning should be 5
designed and implemented to foster learners’ extended engagement, in-game per- 6
formance, and game-based disciplinary knowledge learning and transfer. In this 7
chapter, we review prevalent support features in digital games, prior theoretical and 8
empirical research on scaffolding and support in game-based learning, and the sup- 9
port design conjectures deemed effective. We then share our observations of the 10
obstacles that learners experienced in game-based learning processes when using 11
E-Rebuild, describe the corresponding learning support strategies and features, and 12
report the findings from the iterative testing and refinement of these support fea- 13
tures. Propositions for future research and the design of support for game-based 14
learning are discussed in relation to the current project findings and prior research. 15
6.1 Introduction 16
The role of support for learning in game-based learning contexts cannot be overem- 17
phasized (Wouters & Van Oostendorp, 2013). A closely related and frequently 18
examined construct of support for learning is scaffolding. Though the definition and 19
scope of scaffolding is still inconclusive in the literature, scaffolding originally 20
referred to situations “in which the learner gets assistance or support to perform a 21
task beyond his or her own reach if pursued independently when unassisted” (Pea, 22
2004, p. 430; Wood, Bruner, & Ross, 1976). An adaptive level of support and the 23
fading of the support are implied as intrinsic components of the scaffolding process. 24
In this chapter, we describe our design and research of support for game-based 25
learning in a broader sense that includes both scaffolding with fading (or support as 26
needed) and general learner support without fading. We will use the terms of sup- 27
port and scaffolding interchangeably to include all support features in game-based 28
learning, whether adaptive or general, internal or external. 29
Support for learning in gaming often involves two alternative avenues: internal 30
scaffolding as the innate design feature of the game (or tool−/material-mediated 31

https://doi.org/10.1007/978-3-030-04339-1_6
122 6 Designing Dynamic Support for Game-Based Learning
32 support) and external scaffolding as part of the game-based learning experience or

33 activity (or human agent-mediated support). Research on the desirable implementa-
34 tion context and the corresponding relative effectiveness of the internal tool-
35 mediated learning support or the external learning support with human agency is
36 lacking. It remains inconclusive as to what, when, and how support for learning
37 should be designed and implemented to foster learners’ extended engagement, in-
38 game performance, and game-based disciplinary knowledge learning and transfer.
39 In this chapter, we review prevalent support features in digital games, prior theoreti-
40 cal and empirical research on scaffolding and support in game-based learning, and
41 the support design conjectures deemed effective. We then share our observations of
42 the obstacles that learners experienced in game-based learning processes when
43 using E-Rebuild, describe the corresponding learning support strategies and fea-
44 tures, and report the findings from the iterative testing and refinement of these sup-
45 port features. Propositions for future research and the design of support for
46 game-based learning are discussed in relation to the current project findings and
47 prior research.
48 6.2 P
rior Design and Research on Support in Game-Based
49 Learning
50 6.2.1 Support in Digital Games
51 Support features found in digital games are not necessarily designed for content
52 learning purposes but focus on teaching novice players the flow of a game or the
53 game mechanics (i.e., basic actions, rules, and controls) while providing an optimal
54 level of challenge to support game engagement or flow (Chen, 2007). A common
55 design is a non-interactive in-game tutorial that provides a static introduction of
56 fundamentals, though they are frequently ignored by players who generally have
57 short attention spans and are not looking to digest masses of information. To aid or
58 motivate the tutorial processing, the designer can reward the action of attending to
59 the tutorial with achievement points. An engaging and active substitute of the in-
60 game tutorial is to design an assortment of training levels in which the player starts
61 in a “noob cave” (or fail-proof) game level and then experience or learn key game
62 controls or actions one by one via and during gameplay. Rather than providing
63 explicit instructions, certain games (e.g., the game of Plants vs Zombies) try to use
64 universal, basic rules of gameplay or a fairly standardized user interface, assuming
65 the player has some familiarity with them and do not have to relearn everything.
66 They also encourage gameplay as active self-experimentation.
67 To create an optimal challenge, a frequent design strategy is to progressively
68 increase challenge or the complexity level by increasing the variables to be man-
69 aged in a problem scenario (Bos, 2001). In the game SimCity, for example, the
70 designer associates access to new game mechanics with the progression curve and
71 gradually increases the number of changeable task parameters or rules through lev-
6.2 Prior Design and Research on Support in Game-Based Learning 123
els. Another exemplary support feature is an in-game “technology tree” (e.g., in the 72
game of Civilization) that limits the players’ access to advanced items or tools until 73
they are able to master the usage of the basic mechanics. 74
Apart from internal game support, game community mechanisms, such as “let’s 75
play” gaming videos on YouTube and community support via forums, conferences, 76
and blogs, have provided both learning and socioemotional support by peers and 77
developers. Prior research (e.g., Ho & Huang, 2009; Steinkuehler, 2006) on online 78
game communities has reported that online communities for gamers have reinforced 79
not only participatory culture but also transactional and social learning processes. 80
Because the game designers’ focus is on preventing frustration and reinforcing 81
enjoyment more than the effectiveness of learning embedded domain-specific 82
knowledge, the scaffolding or support approaches in digital games may have the 83
potential to act as learning barriers (Sun, Wang, & Chan, 2011). For example, instant 84
feedback and demonstration scaffolding that aim to reduce frustration levels can 85
reduce the potential for experimentation and failure-driven learning, create overreli- 86
ance on system prompts without internalizing knowledge acquired from interac- 87
tions with the game system, and hence undermine learning if overused. The 88
mechanism that supports a sense of playfulness could also be in conflict with the 89
development of purposiveness and the effort contributed to “problematizing” the 90
game challenges with respect to academic disciplines or subjects (Reiser, 2004). 91
6.2.2 Support in Game-Based Learning 92
Games are a complex learning environment in which players—especially novices— 93

can be overwhelmed by the abundance of information to be processed, the dynamics 94
of the game world, and the complexity of the task and consequently deterred from 95
activities that foster learning (Ke, 2016; Wouters & Van Oostendorp, 2013). 96
Moreover, the dynamic and instant linking between a player’s action and changes in 97
the game world may lead to trial-and-error learning in which knowledge remains 98
intuitive without being explicated or conceptualized (Leemkuil & de Jong, 2011). 99
Hence it is warranted to design supports in game-based learning that facilitate not 100
only the selection and processing of relevant information but also an active organi- 101
zation and conceptualization of intuitive insights for substantive knowledge 102
development. 103
During the past decade, increased research has examined the design features that 104
promote game-based learning, especially learning or instructional support features. 105
Designing supports or scaffolds to fit the game-based task properties as well as the 106
associated control interfaces is a critical design issue (Soloway, Guzdial, & Hay, 107
1994). Though empirical studies examining particular game-based learning support 108
features have increased in the past decade, research on a systematic framework of 109
game-based learning support design is still lacking. The literature has, however, 110
shed light on a design framework for computer-assisted learning support in general. 111
Wood, Bruner, and Ross (1976) initially specified scaffolding as the processes of 112
113 (attention) recruitment, reduction in degrees of freedom for the problem at hand,
114 direction maintenance, marking critical features, frustration control, and demon-
115 stration (or modeling). Building on such a perspective, Pea (2004) and Reiser (2004)
116 observed that scaffolding serves three main major purposes: (a) channeling and
117 focusing (or task structuring), such as reducing the degrees of freedom for the task
118 at hand by providing constraints that increase the likelihood of a learner’s effective
119 action and recruiting and focusing learners’ attention by marking relevant task fea-
120 tures in what is otherwise a complex stimulus field; (b) modeling, such as demon-
121 strating a solution to the task, with or without human agency (e.g., modeling by
122 software features or a socially interactive other); and (c) content problematizing—
123 tools or features that can shape a learner’s performance and understanding of the
124 task in terms of key disciplinary content and strategies while problematizing the
125 disciplinary content with the task at hand. Quintana and his colleagues (2004) then
126 tried defining a scaffolding design framework for computer-assisted scientific learn-
127 ing environments via a theoretical review and a synthetic analysis of exemplary
128 courseware. They classified a list of design heuristics and techniques of scaffolding
129 based on the basic problem-solving processes to facilitate, such as sensemaking (or
130 problem and content representation), process management (or task structuring along
131 with modeling), and articulation and reflection (for content problematizing). These
132 earlier frameworks of scaffolding, along with the previous empirical studies on spe-
133 cific game-based learning support features, have all informed our design of learning
134 support features in E-Rebuild.
135 Wouters and Van Oostendorp (2013) conducted a meta-analytic review of 29
136 studies on the effects of instructional support in game-based learning. They reported
137 that players benefit from instructional support for knowledge (d = 0.33, I < 0.001),
138 skills (d = 0.62, I < 0.001), and in-game performance (d = 0.19, p < 0.001). Support
139 features that facilitate learners in attending to and selecting relevant information
140 (d = 0.46, p < 0.001) are more effective than the ones that stimulate organizing and
141 integrating new information (d = 0.14, p < 0.01). Especially, modeling (or explica-
142 tion of the problem-solving procedure), multimodality, personalization, and feed-
143 back are effective techniques to support learners in selecting relevant information.
144 Their analysis indicated that the most effective instructional support for information
145 organization and integration is reflection (d = 0.29), in which learners are explicitly
146 asked to think about their actions or decisions. Instructional support that less explic-
147 itly stimulates the organization and integration of information, such as narrative
148 elements and collaboration, seems to be less effective (d = 0.11 and d = 0.14, respec-
149 tively). They also reported that the effectiveness of instructional support in game-
150 based learning was significant for elementary school and college/university students
151 (d = 0.19, p < 0.05; d = 0.41, p < 0.001, respectively), but not significant for middle
152 or high school students.
153 The recent work of Kao, Chiang, and Sun (2017) on scaffolding in game-based
154 learning environments reported that marking critical features facilitated concep-
155 tual knowledge acquisition better than demonstration, and the two scaffolding
156 types facilitated different dimensions (sensitivity and flexibility) of design cre-
6.2 Prior Design and Research on Support in Game-Based Learning 125
ativity. Mayer and Johnson (2010a, b) reported that people who played the circuit 157
game learned faster and were better able to transfer to new problems if they 158
received in-game guidance in the form of non-interruptive self-explanation (i.e., 159
selecting a reason for each action) or feedback (i.e., being shown the reason for 160
the correct action). They concluded that the educational impact of a game can be 161
substantially improved by incorporating support features aimed at guiding the 162
learner’s cognitive processing during playing. Their conclusion was supported 163
and extended by a later study by O’Neil et al. (2014) who found that self-expla- 164
nation prompts aimed at helping game-based learners make connections between 165
game variables and disciplinary concepts (or content problematizing) are espe- 166
cially effective. 167
Other studies have similarly examined external supports for game-based learn- 168
ing. For example, Barzilai and Blau (2014) found that conceptual external scaf- 169
folds provided before gameplay led to better problem-solving (in comparison 170
with the play-only condition or external scaffolds provided after gameplay) but 171
lowered perceived learning. Presenting external scaffolds may have conceptual- 172
ized learners’ understandings of the game by connecting them to disciplinary 173
knowledge, but its effectiveness is moderated by the timing of the scaffolding. 174
Tsai et al. (2013) examined both teacher-initiated content presentation before 175
gameplay and student-controlled in-game question prompts as game-based learn- 176
ing scaffolds. They reported that scaffolding before and during gameplay pro- 177
moted knowledge test performance better than scaffolding only during gameplay 178
and non-scaffolding conditions. Promisingly, the embedding of background con- 179
tent objects or question prompts did not influence students’ perceptions of their 180
gaming experiences. Chen and Law (2016) examined and compared two types of 181
external supports for game-based learning: nonadaptive question prompts pro- 182
vided after gameplay (called hard scaffolds) and peer discussions during collab- 183
orative gameplay (called soft or dynamic scaffolds). They found a significant 184
positive effect of both external supports on learning performance but a negative 185
impact on motivation (competence, autonomy, and interest). They also found that 186
after-gameplay question prompts can reinforce the effectiveness of collaboration 187
for game-based learning. 188
In summary, the previous study findings suggest that the content of the scaffolds, 189
as well as the timing of their provision, should be carefully designed according to 190
the game features to achieve specific instructional purposes. Supporting the argu- 191
ments of Wood et al. (1976) and Pea (2004), in-game learning scaffolds that high- 192
light channeling, focusing, and content problematizing while being non-interruptive 193
are found effective. External scaffolds, such as collaboration and question prompts 194
aimed to explicate knowledge development, can be effective, but their effectiveness 195
will be moderated by the timing of the scaffolds. 196
197 6.3 D
esign Conjectures and Observed Challenges in Game-
198 Based Learning
199 The design of game-based learning in E-Rebuild, as described in the previous chap-
200 ters, assumes that core game actions will necessitate transactional and problem-
201 based learning in which players interact with dynamically represented elements of
202 a building-themed math problem, select and coordinate related information embod-
203 ied in game objects, plan and experiment with the problem-solving actions in an
204 interactive game world, and comprehend the task (with its variables and relationship
205 structure) in terms of key disciplinary content and strategies, thus engaging in math-
206 ematical conceptualization and problem-solving. These design conjectures imply
207 multiple expectations on the game-based learning processes—attentiveness, purpo-
208 siveness, and need for cognition of a problem-based or discovery learner, an
209 acquired motive or sense of legitimacy toward “content problematizing,” as well as
210 a prerequisite resource base for performing entry-level tasks (e.g., the basic math
211 content knowledge and skills for building-themed problem-solving).
212 Yet heterogeneity in the school student group and the complex nature of game-
213 based problem-solving induce observed participation and outcome variability in
214 which not all participants respond to game-based learning tasks in an expected or
215 favorable way. In particular, we found that many students lack training in compre-
216 hending and representing the mathematical relationships in a multistep context
217 problem, need help to connect and structure distributed task information embodied
218 in game objects and game actions, and do not automatically engage in the cognitive
219 or affective processes of game-based learning. Some of them also lack entry-level
220 conceptual knowledge or general understanding of design problem-solving.
221 In the initial prototype of E-Rebuild, we tried presenting a multistep, building-
222 themed math problem in its lifelike form via a single game level. A mission descrip-
223 tion (or design problem statement) in the opening task panel worked as an
224 overarching task narrative. Designed as an open-ended learning environment, the
225 game world was a simulated natural landscape in which building items (or materi-
226 als), a potential construction site, and the structure to be built (with its criterion
227 properties) were situated but not explicitly denoted. Relevant cueing was presented
228 within the opening task narrative (e.g., “those shipping containers should be of
229 help”) or as learner-controlled, displayable visual aids (e.g., a “historical view” por-
230 trays the structure to be rebuilt in the pre-quake landscape, via a 2D slideshow as
231 well as a 3D emulation of the pre-quake structure and landscape). These design
232 efforts were intended to highlight aspects of mystery and puzzle in gameplay and
233 game-based, realistic math problem-solving.
234 Frequently, however, participants were not fully processing or digesting a written
235 task narrative at the beginning of gameplay, even when the narrative was concise
236 and evocative. They appeared to treat it as an affective plot opening rather than a
237 mental riddle containing the critical task message and hence lacked attentiveness to
238 delineate the embedded problem goal and variable information. The open-ended
239 game world also reminded some players of commercial building games that they
6.3 Design Conjectures and Observed Challenges in Game-Based Learning 127
played (e.g., Minecraft), who then intuitively transferred the related gaming actions 240
or strategies (e.g., exploring the game world freely and building items at one’s own 241
will and standard) to the current game. As observed, they were often wandering 242
around the 3D landscape, clicking around, and picking up collectible objects ran- 243
domly. They then assembled the collected items casually, without an understanding 244
of the building goal/problem and the related design criteria. They became puzzled 245
upon receiving the level-failed message and asked, “What are we supposed to do?” 246
It was only after multiple failed trials would they realize that their old gaming habits 247
did not work and hence show more attentiveness in reading the task narrative and 248
other in-game cues. Other learning habits or preferences, such as the need for 249
explicit, step-by-step procedural guidance along with the lack of appreciation of 250
independent puzzle solving (or a low need for cognition), made them easily frus- 251
trated from game bottlenecks and frequently asking for instant help from peers or a 252
facilitator, “What should I do now?” Instead of looking to maximize “hard fun,” 253
these players wanted to pass the level and end the problem-solving experience as 254
soon as possible. 255
We also observed that players tended to lack persistent effort in task structuring 256
(i.e., mapping sub-goals and planning steps beforehand). For example, we expected 257
that a player would proceed with a logical process of site surveying, planning the 258
structure to be built (in terms of the size, shape, and position), collecting/trading 259
items needed, building as designed, and allocating inhabitants to the compartments 260
AU1 of the structure. Yet most players were involved in intuitive, trial-and-error game 261
actions that lacked system thinking; instead, building instantly with only items at 262
hand while searching more items as needed, building a random structure, finding it 263
impossible to allocate all inhabitants, rearranging/rebuilding the structure, realloca- 264
tion (potentially with repeated mistakes), and so on. This unsystematic problem- 265
solving process, to some extent, generated failure-driven, reflective understanding 266
about the presence of a mathematical relationship among task variables. It was, 267
however, non-mathematical and thus made it difficult for the players to complete 268
sufficient levels within a class session. To assist these intuitive-thinking players, we 269
provided a short list of sub-tasks (or marking the task structure) on the task panel. 270
This design feature did significantly improve the players’ in-game performance 271
(e.g., the number of levels completed within a session) but reduced the opportunity 272
for the participants to practice representation for multistep problem-solving (Ke, 273
2007). This finding is consistent with prior research which found the big challenge 274
for designing game-based learning support is to find a balance between supportive 275
tool availability and encouraging learners to engage in discovery learning and 276
accept some level of frustration from game bottlenecks (Sun et al., 2011; Yelland & 277
Masters, 2007). 278
E-Rebuild learners also experienced a learning curve in coordinating information 279
distributed across objects and actions in the 3D virtual world. Situating a mathemat- 280
ical problem in a 3D world, representing related problem variables as interactive 281
objects or action feedbacks that are distributed across game space and time, and 282
making one actively search and connect (rather than being provided with) the 283
embedded information were generally novel practices to middle-school students. In 284
285 the initial E-Rebuild gaming sessions, we commonly observed that participants
286 lacked purposiveness in storing and using information obtained during the previous
287 game object or action to guide their next move. Even when sub-tasks were marked
288 explicitly, participants tended to treat them as separate assignments rather than part
289 of a holistic problem. Based on think-aloud protocols and interviews, participants
290 generally lacked a solid understanding of the systematic nature of mathematics—
291 the need to work on a higher logical plane in problem-solving situations—and
292 understanding which and why mathematical relationships and ideas are useful in a
293 particular context for problem-solving.
294 6.4 Pros and Cons of Learning Support Design Strategies
295 Driven by both the literature on game-based learning support and the observed
296 learning challenges in E-Rebuild, we experimented with multiple learning support
297 design strategies: (a) channeling learners’ attention toward the final target task by
298 displaying sub-tasks in structure and constraining the space exploration to maintain
299 directedness of the learner’s activity toward task achievement; (b) presenting/
300 sequencing game tasks (or levels) with gradually increased degrees of freedom, by
301 reducing the numbers of variables in the task structure involved in the initial prob-
302 lem scenarios; (c) modeling or scaffolding the representation of mathematical rela-
303 tionships in a task with interactive step-by-step prompts, an in-game scratch pad,
304 multimodal cues, and feedback; and (d) externally supporting content problematiz-
305 ing by encouraging learning-constructive game talk among peers and external help
306 as needed from a facilitator or teacher. The pros and cons of these learning support
307 design strategies, with the illustration of our design experiment findings, are
308 described in the following section.
309 6.4.1 Structuring and Sequencing of Game Tasks
310 To assist players in mapping and systematically planning problem-solving steps, we

311 have experimented with alternative ways of chunking and structuring a game task.
312 One method involves chunking and sequencing sub-goals (or componential steps)
313 of a composite task, while another involves chunking and then gradually increasing
314 the set of influencing variables of the problem/task.
315 Sub-Goal Chunking and Structuring In our earlier prototype, we tried the first
316 chunking/structuring strategy by evaluating each sub-goal as a componential task,
317 sequencing them linearly as a succession of tasks, and showing only the “instant”
318 sub-goals with a task completion progress bar (Fig. 6.1a). The next sub-goal or task
319 (e.g., allocation) was locked until the previous one (e.g., building) was completed.
320 This design arrangement highlighted a “step-by-step” guidance on task structuring.
6.4 Pros and Cons of Learning Support Design Strategies 129
Fig. 6.1 (a-b) Marking and sequencing of sub-goals
Such an arrangement has obviously fostered in-game task performance and on-task 321
time, in comparison with the initial design that presented a multistep task as it is. 322
Yet we also observed that the explicit and linear sub-goal structuring reduced the 323
opportunity for the participants to proactively reflect on and refine their under- 324
standing of mathematical relationships underlying the task structuring. For exam- 325
ple, in the episode depicted in Fig. 6.1, the building problem (using multiple 326
containers to build a shelter resembling the pre-quake home and allocating all sur- 327
viving families to containers) includes two puzzles: (1) the number of shipping 328
containers needed and (2) the way the structure should be composed. When inter- 329
acting with the initial version, participants were frequently found transferring 330
between the two steps of “rebuilding” and “sheltering all families” to self-check 331
and refine the answer to the first puzzle. Yet under the linear and chunked task 332
structure, both puzzles had to be solved at the building step, and the participants 333
lost meaningful action feedback (e.g., whether all families get sheltered). Hence 334
more participants started trial-and-error gameplay: building with all containers, let- 335
ting the computer or game check the built structure, and reducing one container at 336
each new trial until they passed this step. 337
In a later prototype, we presented all sub-goals of the composite game task as a 338
to-do list (Fig. 6.1b), with which the player can check off each completed sub-task 339
at a nonlinear sequence. All sub-goals or componential tasks are parallel and not 340
prerequisite of each other. The flexible sub-goal list helped to channel the players’ 341
attention while allowing them to navigate and convert between sub-goals (or tasks). 342
343 As the gaming behavior analysis indicated, the nonlinear sub-goal chunking, in
344 comparison with the linear one, was associated with more enactments of content
345 learning engagement (e.g., information processing, calculation, knowledge applica-
346 tion) and generic cognitive engagement (e.g., planning, evaluation, refining).
347 However, providing the players with a specified sub-goal list has reduced the neces-
348 sity for learners to proactively construct and experiment with personally meaningful
349 representations of a multistep math context problem. They thus missed the opportu-
350 nity to practice problem identification with task structuring, a critical element for
351 realistic math problem-solving.
352 Variables Chunking and Escalation To enable an exploration or experience of
353 the systematic nature of mathematical relationships in a problem while scaffold task
354 structuring, we tried examining the number or structure of problem variables as an
355 additional salient entity for game level organization and sequencing. Specifically,
356 we designed and sequenced game tasks with a gradually increasing number of influ-
357 encing variables (or degrees of freedom) involved in the underlying mathematical
358 relationships. In a classroom building level of the School episode (Fig. 6.2a–b), for
359 example, the problem involves multiple variables: unit classroom/container space,
360 unit space need of each student of each subject, the total space/area needed, and the
361 positioning (and distance) of classrooms. The logged failed attempts, the frequency
362 of help requested, and the observed frustration level of participants were all higher
363 than expected in the original version (Fig. 6.2a). We therefore chunked the original
364 level into an easier version (Fig. 6.2b) where the positioning/distance variable was
365 temporarily removed. In the next level, a more composite version of the task was
366 then presented with the full set of the variables. The revision increased participants’
367 task performance and content-engaging actions.
Fig. 6.2 (a, b) Chunking and sequencing by task variables

Fig. 6.3 The plan view of the school set design
Constraining Game Space and Highlighting Information Points of Interest We 368

additionally found that constraining the game space helped to reduce the task- 369
irrelevant locale exploration acts and channel participants’ attention toward prob- 370
lem information processing and selection. Take the aforementioned School episode 371
as an example; the earlier version presented a full replication of an urban school 372
setting (Fig. 6.3), designed by the architect consultant on the team. Proceeding from 373
the previous Desert episode to this lifelike School episode, participants (at a local 374
school district in Florida) all demonstrated a high level of excitement with the rich- 375
ness and attractiveness of the 3D game world. They became immersed in the envi- 376
ronmental story portrayed by the game space: navigating and checking every 377
structure and background object of the set and asking the facilitator whether they 378
could drive a school bus or how to enter a locked townhouse. They spent more time 379
exploring the task-irrelevant game space than fully processing the task narrative and 380
related information objects. Due to this observation and the need to downsize the 381
game file to run on the local school infrastructure, we reduced the magnitude of the 382
school set and background structures/objects. To better focus the players’ attention 383
Fig. 6.4 The school stadium level—starting point
384 on the salient problem variables (or the embodied game objects), we purposefully
385 made the information point of interest—the object that is critical for the problem
386 identification—as the starting point of gameplay at each level. For instance, in a
387 school stadium building level (Fig. 6.4), participants had to analyze the structure of
388 the stadium’s bench seating and find a way to build it in an efficient way. The prob-
389 lem requires the player to decompose and compose geometric shapes (e.g., a trap-
390 ezoidal prism as the base and a rectangular prism as the seat) and join and copy/
391 paste small cuboids of standard or varied dimensions (e.g., bricks) to construct the
392 base and seat of the stadium bench. This task, as observed, was novel to our middle-
393 school participants who struggled in analyzing the composition of the structure,
394 usually spending more than 30 minutes in a single attempt or trial, and frequently
395 requesting hands-on help with the building process. Specifically, they did not fully
396 inspect the bench to notice its hollow rear. To correct this misapprehension, we set
397 an isometric view of the bench’s rear (instead of the elevation view of its front side,
398 as in the earlier version) as the starting point of the level. We also colored the base
399 and seat section of the bench differently to highlight its geometric composition. All
400 of these design revisions with the game space have clearly reduced the off-task time
401 and actions by the participants.
Fig. 6.5 (a, b) Scaffolds of mathematical relationship representation in iconic and symbolic
formats
6.4.2 I n-Game Scaffolding of Mathematical Relationship 402
Representation 403
To scaffold players’ comprehension and representation of the mathematical rela- 404

tionships in a game task, we tried adding step-by-step, interactive prompts, pre- 405
sented in both pictorial (or iconic) and symbolic formats, in the Help panel (see 406
Fig. 6.5). An empirical investigation of this scaffolding feature (Lee, 2016; Lee & 407
Ke, 2016) found a significant and positive impact of using this representation scaf- 408
fold, especially in symbolic form, on learners’ conceptual and procedural mathe- 409
matics knowledge development via gameplay. Yet the infield observations indicated 410
that players would not necessarily attend to or process the mathematical scaffolds 411
unless they were associated with explicit game rewards. 412
Tina got stuck on the numerical expression and calculation of a family’s space 413
need and started to use a trial-and-error strategy with the prompt—entering random 414
values to the text area until she guessed the right one. The scaffolding window pre- 415
sented instant feedback on every entry (e.g., reddening the wrong entry without 416
proceeding to the next screen), aiding the trial-and-error practice. Proceeding to the 417
next prompt, Tina kept guessing on a related math variable—the total space needed. 418
It appeared that she did not understand or remember the ratio statement and the vari- 419
able of family space processed/calculated during the previous step. Her neighbor 420
asked, “How did you figure out what (number) goes here (a variable in the equa- 421
tion)?” She answered, “I was just typing numbers.” 422
As this observation shows, certain participants failed to associate the interactive 423
scaffolds on numerical expressions with the game task at hand and deemed these 424
scaffolds as boring hindrances during gameplay. In contrast, other participants, 425
426 driven by multiple failed trials of the game task, were cognitively engaged in inter-
427 acting with the scaffolds.
428 Jeremy got stuck in the family allocation task and failed multiple times. He
429 started to check the help panel. It appeared that this was his first time reading the
430 scaffolding section seriously. He completed the first three ratio prompts smoothly.
431 He entered correct variable values in the equation for calculating the space needed
432 for a 1-adult-2-child family, but did not calculate the total correctly until the third
433 trial, indicating a lack of competency in interpreting and calculating the mathemati-
434 cal expression. He got stuck with the equation for a 2-adult-1-child family in the
435 next step; he entered wrong values for the space needs of each adult and child, lack-
436 ing a solid understanding of the math concept of unit. He went back to the previous
437 step and reviewed the equation of the 1-adult-2-child family. After fully compre-
438 hending that mathematical expression, he then reworked on the equation for the
439 2-adult-1-child family. This time he entered correct values for both the unit and total
440 space needs.
441 In the above example, the participant had developed a better understanding of the
442 unit concept and the mathematical equation representing the space needs of varied
443 families. As exemplified, in-game scaffolds assisted participants’ encoding of the
444 math problem using mathematical expressions or notations, but only when they
445 were processed mindfully. Therefore, it is important to design the in-game scaffold
446 of mathematical representation in a way that the players will perceive the scaffold
447 as a sensible investment of cognitive effort and an essential support for their game-
448 play or game-based problem-solving process.
Fig. 6.6 An in-game

scratch pad
In-Game Scratch Pad In E-Rebuild, we designed an in-game scratch pad (Fig. 6.6) 449
as a virtual calculator and a virtual worksheet with which the player can write, plan, 450
and record their problem-solving steps and calculation processes. Specifically, 451
when conducting a site survey using the measurement tool, the player could conve- 452
niently select and record measurements to the scratch pad for quick numerical cal- 453
culations. The infield testing indicated that players frequently used it as a calculator 454
and a notepad for measurement recording, yet few used it to scribe or plan problem- 455
solving steps. A critical reason for this finding is that electronically typing mathe- 456
matical expressions and equations is effortful for middle-school students. In 457
addition, the players could not map or encode the mathematical relationships via 458
freehand diagrams or visuals, on the scratch pad. Hence instead of using the virtual 459
scratch pad, the players preferred to use the offline paper worksheet to aid their 460
problem-solving and calculation processes. However, using an offline worksheet 461
could be interruptive to the players’ gameplay flow and not quite compatible with 462
the desk space that was already packed with the computer, keyboard, and mouse. 463
Consequently, participants tended to either work on a small sheet of paper in 464
between the equipment or choose to bypass the problem-solving scribing or map- 465
ping processes. Creating a user-friendly, efficient, and multimodal in-game scribing 466
tool is a continuing design challenge in E-Rebuild. 467
Multimodal Cues and Feedback To recruit and maintain the players’ attention on 468
the salient objects and features in the task structure and facilitate the reflection and 469
purposive refining of problem-solving actions, we designed in-game scaffolds also 470
as tooltips, action-specific feedback, in-game tools, and end-of-level badges. These 471
in-game scaffolds are multimodal, in line with the game mechanics and game world, 472
and presented in response to a game act or at the end of a game level. 473
In the original prototype of E-Rebuild, we tried putting all salient problem infor- 474
mation, including core variables and their features, into the task narrative being 475
portrayed on the Episode panel when one entered a game level and ever-present as 476
a background information object. However, we found that on average, participants 477
spent less than 1 minute reading the task narrative, with some simply skipping it. We 478
then tried presenting salient information as a tooltip associated with each game 479
object or element. The information appears when an anchor was positioned over the 480
related object, as shown in Fig. 6.7. As observed, interacting with diverse informa- 481
tion objects in the game world had involved participants in an in situ and active 482
investigation of the related problem variables and helped them to comprehend the 483
problem through hands-on discoveries. 484
During the design experiments, we found that action-specific cues, in compari- 485
son with ever-present cues, are better received and processed. For example, in the 486
trading action, we used to present items’ prices via a background bulletin board that 487
replicates the format of a math ratio table (Fig. 6.8a). We observed that middle- 488
school participants frequently ignored the board in the store house. They would 489
make a random estimate of an item’s price, make an unreasonably high offer, pur- 490
chase certain items, run out of resource credits during the task, and thus needed to 491
Fig. 6.7 Salient variable

information presented as a
tooltip
492 retry the task. Only after multiple failed attempts in the trading action would they be
493 willing to carefully read the price bulletin board. In response, we added a tooltip
494 presentation of building items’ prices in the inventory, which has reduced trial-and-
495 error processes and promoted numerical calculation behaviors by participants dur-
496 ing the training action.
497 A similar pattern was also observed with the action-specific feedback. For the
498 allocation action in E-Rebuild, we designed an item-allocating panel (Fig. 6.8b).
499 During the infield testing, participants complained that the panel was blocking the
500 game world, and they could not see whether and what item (e.g., a specific type of
501 family) was added or not. Correspondingly, they were found randomly clicking on
502 the adding or reducing buttons without purposive planning or mathematical reason-
503 ing. As a result, we changed the allocating interface to enable action-specific, imme-
504 diate visual feedback: a player can drag items from the inventory to the target space,
505 with the items either appearing in the space (if applicable) or reverting back to the
506 inventory (in a spaced-out situation). We also added a reddened alert at the bottom
507 of the screen cueing them about the reason for each unfitting occasion. The provi-
508 sion of action-specific, intuitive feedback has again fostered participants’ mathe-
509 matical reasoning and calculation behaviors during the game action.
510 To activate students’ reflection on their actions, we tried to present written, infor-
511 mative feedback in a game performance summary at the end of each game level. But
512 again, few participants carefully read the summary; most only skimmed through it
513 before proceeding to the next game level. Although participants did attend to the
514 three game credits presented on the left bottom corner of the screen, many of them
515 did not fully comprehend the relationship between each credit and their specific
516 gaming behaviors. To encourage action-based reflective learning and mathematical
517 reasoning during gameplay, we have instead designed and presented a badge system
518 on the post-level summary (Fig. 6.9). Each badge is designed and presented to
Fig. 6.8 (a, b) Price bulletin board for the training action and the item-allocating panel
encourage a player to improve the efficiency and accuracy in their game-based 519
problem-solving strategies and solutions and hence motivate a purposive usage of 520
mathematical reasoning and knowledge in the future game levels. 521
2D/3D Visuals for Alternative Processing and Comprehension An example dem- 522
onstrating the benefit of multimodal in-game scaffolds for problem representation 523
was the pairing between a 2D visual sketch and a 3D model of the target architec- 524
tural structure in E-Rebuild (Fig. 6.10). The former was presented as a slideshow, 525
Fig. 6.9 Post-level game

performance badges
526 while the latter was presented as a historical view of the structure and construction
527 site with which the player could actively inspect and survey. When interviewed,
528 participants generally reported that they referred to both forms of the visual-spatial
529 representation during gameplay, which was confirmed by the infield observation of
530 their gaming behaviors. During the building action, participants frequently shifted
531 between reviewing the static, visual depiction (with numerical marking) of the tar-
532 get structure and exploring/measuring the structure model situated in the simulated
533 construction site. This pattern confirmed our speculation that such a multimodal
534 representation scaffold enabled learners to comprehend static visuals while interact-
535 ing with a maneuverable object or model, thus sufficing both visual and kinesthetic
536 ways of information processing.
537 In a recent study on the patterns of scaffolding usage of participants of E-Rebuild
538 (Dai, Ke, & Pan, 2017), we found that the help panel was the most frequently used
539 in-game scaffold and was used for the longest time, followed by tooltips, prompts,
540 and summary screen. However, the time spent on these in-game scaffolds was gen-
541 erally short, ranging from 6.85 to 44.65 s per 45-min gaming session. The frequency
542 of using in-game scaffolds has a positive correlation (r = 0.20, p < 0.05) with the
543 frequency of learning engagement.
Fig. 6.10 2D floorplan slideshow and 3D historical view of the structure
6.4.3 External Scaffolding 544
In addition to the in-game scaffolding, we also encouraged learning-constructive 545

game talk among peers and external help as needed from a facilitator during partici- 546
pants’ gaming sessions. Even though E-Rebuild was developed as a single-player 547
game, participants voluntarily partnered up with their neighbors during gameplay, 548
updating or showing off their gaming progress with each other, sharing tips or strat- 549
egies, or venting frustration. Though the game talk was not fully mathematics rele- 550
vant, it acted as an external support for game task engagement. Via the game talk, 551
student participants frequently sought help and affective support from each other. 552
They also relied on the external scaffolding offered by a trained facilitator. The 553
facilitator’s scaffolding in the E-Rebuild program included question-answering on 554
the game mechanics, modeling of the 3D object maneuvering procedure, cueing on 555
the information to be processed or the potential way to unblocking a bottleneck, and 556
prompting for action-based reflective thinking. Notably, we found that external 557
scaffolding by the facilitator was especially beneficial for the participants who were 558
underachieving in mathematics. There was a positive association between the facili- 559
tator scaffolding usage and these students’ post-gaming math knowledge test per- 560
formance. Yet these students, in comparison with their peers, were significantly 561
more passive in seeking external help. We thus requested the facilitators to provide 562
proactive and recurrent scaffolding to the underachieving students, though it was 563
difficult in cases when multiple students required help. 564
Our study on the scaffolding usage in E-Rebuild (Dai, Ke, & Pan, 2017) indi- 565
cated that student participants generally sought a similar number of human facilita- 566
tor scaffolds and in-game scaffolds, but they spent much longer time with human 567
facilitators. In addition, participants received much more scaffolding, in both fre- 568
quency and time spent, from adult facilitators than from their peers. There was a 569
significant correlation between facilitator scaffolding and learner’s engagement 570
behaviors (r = 0.26, p < 0.05). 571
572 6.5 C
onclusive Design Insights on Game-Based Learning
573 Support
574 Our design and research findings in E-Rebuild have highlighted and confirmed a
575 salient claim in the literature on computer-assisted learning support—only when a
576 support is context-aware and intrinsically integrated in the technical environment
577 will the support be effective. In game-based learning, a support should not only be
578 context- or environment-coherent but also action specific. Game-based learners
579 have a short attention span for textual information processing in a dynamic game
580 world and are occupied with in-game problem-solving. Consequently, they tend to
581 interact with learning supports when they are game action- and object-associated,
582 perceived as a legitimate investment (e.g., a worthy solution to block repetitive fail-
583 ures), and presented in a multimodal and non-interruptive way (i.e., not constituting
584 extra information-processing load).
585 Confirming the literature on game-based learning support (Wouters & Van
586 Oostendorp, 2013; Mayer & Johnson, 2010), the current project findings suggest
587 that learning supports that scaffold mathematical relationships (or task structuring)
588 are beneficial to learners’ in-game task performance and learning outcomes.
589 Specifically, chunking and sequencing sub-goals as well as gradually increasing
590 componential variables in the game tasks help to channel the players’ attention and
591 effort toward the overarching task goal and a systematic problem-solving process.
592 However, it remains a design challenge to create an in-game task-structuring tool
593 that enables mathematical diagraming or freehand scribing or a tool that bridges
594 action- and visualization-based qualitative insights to mathematical expressions of
595 the variables with their analytical relationships in an architectural building
596 problem.
597 In the E-Rebuild project, we have found that internal and external scaffolding
598 can complement each other to create a coherent learning support system. Proactive
599 and individualized external support particularly assists underachieving students. Yet
600 individualized learning support via a trained facilitator is demanding in the school
601 settings. A potential solution and future research topic of the E-Rebuild project is to
602 provide in-game adaptive learning support that is driven by the real-time assessment
603 of the cognitive and affective states of the learners during gameplay. For example,
604 the game data-driven assessment of a learner’s math competency and affective states
605 can drive an adaptive selection and sequencing of game levels that supports an opti-
606 mal learning path and level of challenge. Within each game level, the real-time
607 tracking of the learner’s in-game task performance and engagement state will then
608 trigger multimodal support features that are adaptive in the content (e.g., associated
609 with the action at hand or the object being interacted with), timing (e.g., when a
610 threshold value of failed trials or the bottleneck state is reached), and format (e.g.,
611 modeling or prompting for a partial solution). These adaptive learning support strat-
612 egies can augment the current in-game scaffolding features and balance learner
613 autonomy in selecting the preferred support mechanism.
References 141
AU2 References 614

AU3
Bos, N. (2001). What do game designers know about scaffolding? Borrowing SimCity design 615
principles for education. Ann Arbor, MI: University of Michigan and CILT, PlaySpace Project, 616
Retrieved from http://concepts.concord.org/projects/PlaySpace/application/pdf/Learning%20 617
from%20SIMCITY.pdf. 618
Barzilai, S., & Blau, I. (2014). Scaffolding game-based learning: Impact on learning achievements, 619
perceived learning, and game experiences. Computers & Education, 70, 65–79. 620
Chen, C. H., & Law, V. (2016). Scaffolding individual and collaborative game-based learning in 621
learning performance and intrinsic motivation. Computers in Human Behavior, 55, 1201–1212. 622
Chen, J. (2007). Flow in games (and everything else). Communications of the ACM, 50(4), 31–34. 623
Dai, Z., Ke, F., & Pan, Y. (presented 2017, July). How can we help? Examining learning behavior 624
patterns in a math game. Presentation at 2017 Digital Games Research Association Conference, 625
Digital Games Research Association, Melbourne: Australia. 626
Ho, S. H., & Huang, C. H. (2009). Exploring success factors of video game communities in hier- 627
archical linear modeling: The perspectives of members and leaders. Computers in Human 628
Behavior, 25(3), 761–769. 629
Johnson, C. I., & Mayer, R. E. (2010). Applying the self-explanation principle to multimedia 630
learning in a computer-based game-like environment. Computers in Human Behavior, 26(6), 631
1246–1252. 632
Kao, G. Y. M., Chiang, C. H., & Sun, C. T. (2017). Customizing scaffolds for game-based learn- 633
ing in physics: Impacts on knowledge acquisition and game design creativity. Computers & 634
Education, 113, 294–312. 635
Lee, S., & Ke, F. (presented 2016, April). The effects of representation format in problem repre- 636
sentation on qualitative understanding and quantitative proficiency in a learning game con- 637
text. Presentation at American Educational Research Association 2016 Annual Convention, 638
American Educational Research Association, Washington, DC. 639
Lee, S. (2016). The effects of representation format in problem representation on qualitative 640
understanding and quantitative proficiency in a learning game context. (Doctoral dissertation, 641
The Florida State University). 642
Leemkuil, H., & de Jong, T. (2011). Instructional support in games. In S. Tobias & J. D. Fletcher 643
(Eds.), Computer games and instruction (pp. 353–369). Charlotte, NC, US: IAP Information 644
Age Publishing. 645
Mayer, R. E., & Johnson, C. I. (2010). Adding instructional features that promote learning in a 646
game-like environment. Journal of Educational Computing Research, 42(3), 241–265. 647
O’Neil, H. F., Chung, G. K., Kerr, D., Vendlinski, T. P., Buschang, R. E., & Mayer, R. E. (2014). 648
Adding self-explanation prompts to an educational computer game. Computers in Human 649
Behavior, 30, 23–28. 650
Pea, R. D. (2004). The social and technological dimensions of scaffolding and related theoretical 651
concepts for learning, education, and human activity. The Journal of the Learning Sciences, 652
13(3), 423–451. 653
Quintana, C., Reiser, B. J., Davis, E. A., Krajcik, J., Fretz, E., Duncan, R. G., et al. (2004). A scaf- 654
folding design framework for software to support science inquiry. The Journal of the Learning 655
Sciences, 13(3), 337–386. 656
Reiser, B. J. (2004). Scaffolding complex learning: The mechanisms of structuring and problema- 657
tizing student work. The Journal of the Learning Sciences, 13(3), 273–304. 658
Soloway, E., Guzdial, M., & Hay, K. E. (1994). Learner-centered design: The challenge for HCI in 659
the 21st century. Interactions, 1(2), 36–48. 660
Steinkuehler, C. A. (2006). Why game (culture) studies now? Games and Culture, 1(1), 97–102. 661
Sun, C. T., Wang, D. Y., & Chan, H. L. (2011). How digital scaffolds in games direct problem- 662
solving behaviors. Computers & Education, 57(3), 2118–2125. 663
664 Tsai, F. H., Kinzer, C., Hung, K. H., Chen, C. L. A., & Hsu, I. Y. (2013). The importance and use of
665 targeted content knowledge with scaffolding aid in educational simulation games. Interactive
666 Learning Environments, 21(2), 116–128.
667 Wood, D., Bruner, J. S., & Ross, G. (1976). The role of tutoring in problem solving. Journal of
668 Child Psychology and Psychiatry, 17(2), 89–100.
669 Wouters, P., & Van Oostendorp, H. (2013). A meta-analytic review of the role of instructional sup-
670 port in game-based learning. Computers & Education, 60(1), 412–425.
671 Yelland, N., & Masters, J. (2007). Rethinking scaffolding in the information age. Computers &
672 Education, 48(3), 362–382.
Author Queries

AU1 Please check sentence starting “Yet most players…” for
completeness.
AU2 Reference “Johnson & Mayer (2010)” was not cited anywhere
in the text. Please provide in text citation or delete the reference
from the reference list.
AU3 References “Ke, 2016, 2007; Mayer and Johnson (2010a, b)”
are given in the text buit not provided in the reference list.
Please provide.
Chapter 7 1
An Evolving Design Framework 2
for Game-Based Learning Platforms 3
Abstract In this concluding chapter, we discern and discuss an evolving, experiential 4

design framework for game-based learning platforms, by synthesizing the salient 5
design problem-solving events, experiences, and solution-exploration findings 6
reported in the previous chapters. It is not aimed to be prescriptive or exhaustive, but 7
acts as a starting point for specifying the structuring and important concepts of the 8
interdisciplinary design of game-based learning. Results of our phenomenological 9
inquiry assert that game design problem-solving is mainly about problem structuring 10
or transforming an indeterministic problem space to partially limited. The structuring 11
of the interdisciplinary design of game-based learning platforms consists of three 12
norms: (a) transformation of the design goals into functional specifications of the 13
design artifact, (b) coevolution of the problem and solution spaces and exploration and 14
syntheses of partial solution of the subproblems, and (c) use of a common symbol (or 15
representation) system to communicate and focus information while augmenting col- 16
lective memory and processing. 17
7.1 Introduction 18
AU1 Design is innately an ill-structured problem and a phenomenon difficult to model. 19

Design studies on ill-structured design problem-solving, such as those in the fields 20
of architecture and engineering, were mainly aimed to discern a generic design 21
model that is versatile in describing the invariant nature of design problem-solving 22
across the fields (e.g., Brown & Chandrasekaran, 2014; Dorst & Cross, 2001). 23
These generic design models shed light on our examination and understanding of 24
the game design process. But they come short of suggesting tangible solutions or 25
explaining applied reasoning processes for educational game design. 26
In this concluding chapter, we discern and discuss an evolving, experiential 27
design framework for game-based learning platforms, by synthesizing the salient 28
design problem-solving events, experiences, and solution-exploration findings 29
reported in the previous chapters. This framework is aimed to extend and enrich 30
previous generic design models with discipline-specific and functional design 31

https://doi.org/10.1007/978-3-030-04339-1_7
144 7 An Evolving Design Framework for Game-Based Learning Platforms
32 knowledge emerged from our phenomenological inquiry. It is not aimed to be

33 prescriptive or exhaustive but acts as a starting point for specifying the structuring
34 and important concepts of the interdisciplinary design of game-based learning.
35 7.2 Structuring of the Game Design Spaces
36 Game design is characterized by an open-ended or unspecified goal state and an

37 extensive, indeterminate design problem space—in which the transition states (or
38 paths) between the design input and output (or goal) states can be unlimited and
39 their connections are unknown. Results of our phenomenological inquiry assert that
40 game design problem-solving is mainly about problem structuring (Newell &
41 Simon, 1972; Simon, 1973) or transforming an indeterministic problem space to
42 partially limited. The structuring of the interdisciplinary design of game-based
43 learning platforms consists of three norms: (a) transformation of the design goals
44 into functional specifications of the design artifact; (b) coevolution of the problem
45 and solution spaces, including defining the space constraints, decomposing the
46 problem into interconnected components, and exploration and syntheses of partial
47 solution of the subproblems; and (c) use of a common symbol (or representation)
48 system to communicate and focus information while augmenting collective mem-
49 ory and processing.
50 7.2.1 T
ransformation of Interdisciplinary Design Goals
51 into Functional Specifications
52 To formulate the design problem, we have tried to explicate, synthesize, and trans-
53 form the design goals into functional specifications of the design product. Because
54 game design is the experience design, the problem formulation investigation is
55 hence driven by a basic question: What learning-gaming experience will we design?
56 Specifically, a design team has to address the following sequence of questions to
57 specify the mechanics nature of the game-based learning experience:
58 • What composes a salient problem-solving or epistemic practice in the disciplines
59 (e.g., mathematics and architecture)?
60 • What desirable aesthetic experiences—emotional and cognitive responses
61 evoked in the player—are associated with such a practice?
62 • What game objectives, actions, rules or constraints, and environmental storytell-
63 ing will promote the relevant disciplinary practices and aesthetic experiences?
64 • Is designing such an experience applicable considering the technical and prag-
65 matic constraints (e.g., time, budget, resources at hand, and the future implemen-
66 tation setting)?
7.2 Structuring of the Game Design Spaces 145
The above questions have helped to bring to bear interdisciplinary team 67

members’ personal and disciplinary knowledge on the design problem formula- 68
tion and channel their attention and discussion on the core game mechanics and 69
game world design. 70
7.2.2 Coevolution of the Design Problem and Solution Spaces 71
Prior design research observes that creative design is “a matter of developing 72

and refining together both the formulation of a problem and ideas for a solution, 73
with constant iteration of analysis, synthesis and evaluation processes between 74
the two notional design ‘spaces’—problem space and solution space” (Dorst & 75
Cross, 2001, p. 434). Such a design observation got confirmed by our design 76
inquiry findings. Our design chronicle indicates that we start by exploring the 77
problem space and identifying a partial structure (e.g., what is the core game- 78
play?). The partial problem structure then helps us to identify a partial structure 79
of the solution space (e.g., a “building” action along with the 3D movement 80
interface). We examine and use this partial structure to generate more initial 81
ideas for the form of a design concept (such as objectives, potential challenges, 82
and rules associated with a building action), and so extend and develop another 83
partial structure of the solution space (e.g., actions of allocation and collection). 84
These solution space evolutions transfer back to extend the problem space iden- 85
tification (e.g., how will a game task assemble and legitimize game actions in 86
meaningful context?) and, then again, trigger the exploration of the solution 87
space (such as the task design and chunking and the game world development). 88
Ultimately, we create a system of matching problem-solution pairs that are also 89
interconnected with each other. 90
It should be noted that the evolution of the system of matching problem-solu- 91
tion pairs for game design can be inexhaustible. To define constraints or a func- 92
tional boundary to the design space, the interdisciplinary team needs to frequently 93
review and refine the aforementioned functional specifications of the design arti- 94
fact, and prioritize the design structures within the negotiated, goal-driven 95
boundary. 96
A foundational structuring process of the game design space is to decompose 97
(Alexander, 1964) the problem into subproblems that are not strictly encapsulated, 98
with an ongoing monitoring process that looks for interconnections across them. In 99
E-Rebuild, for example, the problem-solution space exploration is decomposed into 100
the following interconnected facets: 101
• Designing game actions and rule sets that necessitate and legitimize the compe- 102
tency performance 103
• Designing an intermediary user interface to enable and motivate both functional 104
and cognitive interactivity 105
106 • Designing and structuring game tasks, including contextual scenarios (the game
107 world or setting), to exemplify actions in a legitimate (or meaningful) context, to
108 scaffold learning progression, and to capture/accumulate performance evidence
109 defined by the competency/evidence models
110 • Designing the game logging structure along with the development of the statisti-
111 cal model for stealth assessment (in this case, Bayesian network)
112 • Designing game-based learning support to enhance theoretical thinking and
113 subject-problematizing opportunities
114 In the later stage of the design space structuring, the sub-solutions for the above
115 subproblems are then aligned and adjusted coherently to create a synthetic design
116 solution for the whole game design problem. For example, in the later design phase
117 of E-Rebuild, game task development and sequencing, game logging structure
118 design, assessment model refinement and validation, and learning support design
119 have merged with each other so that the iterative design and refinement are pro-
120 ceeded with them all rather than a single facet.
121 7.2.3 C
onstruction and Use of a Common Symbol (or
122 Representation) System
123 Using a common symbol system for design concepts among the interdisciplinary
124 design team is imperative for a longitudinal, complicated design problem-solving
125 process. During the E-Rebuild project, we found that a substantial amount of time
126 during the earlier design meetings was spent in decoding, comparing, and coordi-
127 nating differing symbol or representation systems of the design ideas, more than
128 that for the design idea generation. Due to the lack of a common symbol or repre-
129 sentation system of the design knowledge among the interdisciplinary teammates,
130 the collaborative review and refining of the design spaces become even murkier. The
131 use of a common “pattern” language, as we discovered, helps to communicate, fil-
132 ter, and focus information and augment collective memory and processing. The con-
133 struction of such a design language system should actually occur as the
134 commencement of the design problem structuring process.
135 7.3 C
ore Design Concepts for Game-Based Learning
136 Platforms
137 Several core design concepts have emerged as the most dynamic, composite, or
138 impelling constructs in the design spaces of game design. They embody the unique
139 features and challenges that characterize game-based learning platforms and deserve
140 further discussions and a deliberate examination.
7.3 Core Design Concepts for Game-Based Learning Platforms 147
7.3.1 Intrinsic Knowledge Integration 141
Prior research of game-based learning has examined and discussed the importance of 142
integrating knowledge as an internal or intrinsic part of gameplay, such as the design 143
exploration of endogenous fantasy (Habgood, Ainsworth, & Benford, 2005; Malone, 144
1981) and a recent synthesis of major approaches of integrating purposeful learning in 145
gameplay (Ke, 2016). Actually, the nature of the knowledge is as important as the way it 146
is integrated in or structured/enabled by gameplay. But the former receives much less 147
design or research attention. Few game studies examine the nature of game-based knowl- 148
edge to justify why its development warrants gameplay. Besides, an implicit assumption 149
of “embedding” or “integrating” knowledge in gameplay is that the knowledge is infor- 150
mation or static data to be represented or conveyed by multisensory game worlds. 151
In this project, we explore the design of game-based learning as more an experience 152
design, with the aim to convey knowledge in interactions rather than static data. The 153
assumption is that problem-based gameplay enables not only conceptual and proce- 154
dural math knowledge but also structural understanding—understanding of the princi- 155
ples and conceptual relations among mathematical variables and procedures in a 156
contextual math problem. It is also hypothesized that learners will discern new prob- 157
lem-solving methods from the experience, after being provided expressive ways to 158
confront with and test the previously developed methods (e.g., trial and error) and real- 159
izing that they are restricted. These theoretical and design conjectures about learning in 160
gameplay have surely influenced our design exploration. It is important, but beyond the 161
scope of this book, to examine how the nature of knowledge (e.g., general schema ver- 162
sus domain-specific knowledge; sign-mediated theoretical thinking versus empirical 163
thinking) will mediate the design approach of game-based learning platforms. 164
7.3.2 Meaningful Play 165
Although games have been frequently cited as a motivation tool for learning, design 166
research that examines the construct and design of game-based learning motivation 167
is still lacking (Star, Chen, & Dede, 2015). It is important for the game designers to 168
explicate why and how different learners will find a game-based, intentional learn- 169
ing experience interesting (due to intrinsic motivation) or important (due to well- 170
internalized extrinsic motivation). For example, the game-based learning designers 171
need to resolve the conflict between the gameplay that requests considerable sense- 172
making or cognitive interactivity and learners who may lack need for cognition—an 173
inclination toward effortful cognitive activities. 174
Educational game design is to create meaningful, interactive, and challenging 175
worlds involving the user as the conductor of his own intellectual development. In 176
relation to this stance are three design concepts imperative for describing and design- 177
ing game-based learning motivation: autonomy with environmental regulation, 178
meaning in interactivity, and optimal challenge. 179
180 Autonomy balanced with environmental regulation Learner autonomy (e.g.,

181 user choices of the game video/audio setting, speed, and the play mode or path) is a
182 common theme in the discussion of game engagement and usability. But we discov-
183 ered that too much freedom in a game-based learning space is associated with off-
184 task behaviors and the lack of purposiveness and mindfulness for knowledge
185 application and acquisition during gaming. Environmental regulations of learner AU2
186 autonomy, such as rules and incentive structures that prioritize domain-specific
187 methods and tools as well as a game world with a natural boundary, are imperative
188 for game-based cognitive and learning engagement and should.
189 Meaning in interactivity Salen and Zimmerman (2005) discussed meaningful

190 play as an emotional and psychological experience in a game that emerges from the
191 relationship between player action and system outcome, when such a relationship is
192 both discernable and integrated into the larger context of the game. This proposition
193 is adopted and confirmed by our inquiry that the user interface should communicate
194 the result of the game action to the player in a perceivable way while demonstrating
195 the significance of the game action on the play experience both immediately and at
196 a later point in the game. On the other hand, we found that the mechanism of creat-
197 ing meaning in cognitive interactivity (or psychological and intellectual participa-
198 tion between a participant and a system) is much more complicated than that for
199 functional interactivity (or technical usability). For example, players may fail to
200 discern the advantages of using domain-specific methods and tools or associate
201 their significance with the larger context in the game. Design principles or strategies
202 aimed to make cognitive interactivity discernable and integrated in game mechanics
203 and settings warrant deliberation.
204 In the E-Rebuild, we have tried reifying values of cognitive interactivity into
205 gameplay rules and incentive structures (e.g., a game badge system emphasizing the
206 significance of efficiency and thriftiness as the consequence of using mathematical
207 methods or solutions). We have also tried leveling the entry barriers in cognitive
208 interactivity, by chunking and sequential embedding of domain-specific challenges
209 in game levels. We design the game world and the fantasy environment as a context
210 that justifies cognitive interactions. Specifically, discovering and interacting with
211 objects in the game world (e.g., the actions of site surveying and item collection) are
212 intertwined with intellectual participation (e.g., coding and calculating the mathe-
213 matical properties of the site and items). Infield testing and observing the interactiv-
214 ity between players and the coded game system have enabled us to analyze the
215 association between the game task/level features and the creation or reduction of the
216 undesirable shortcut solutions (solutions without domain-specific thinking). In gen-
217 eral, the game tasks should be designed to provide the player expressive ways to
218 confront, test, and compare domain-specific methods or tools with generic ones.
219 Optimal challenge is the most conversed design concept or standard in the previ-
220 ous game studies, whereas it remains the trickiest design challenge due to the high
221 heterogeneity in relation to prior game skills, domain competencies, as well as the
222 general learning ability of the players. It is difficult to gauge the presence of optimal
7.3 Core Design Concepts for Game-Based Learning Platforms 149
challenge until the actual responses of diverse players toward the tasks are iteratively 223
observed. In the E-Rebuild project, we have iteratively retuned the usability of the 224
interaction controls and training levels and emphasized sequential embedding of 225
novel challenges and mechanics in game levels to present a positive slope of the 226
learning curve. 227
On the other hand, an assumption of game-based knowledge development is 228
to reinforce a scientific attitude, such as the persistence toward frustration from 229
bottleneck as well as independent problem-solving. Design efforts to increasing 230
the players’ game engagement level by chunking and reducing the complexity of 231
a game task, or the arrangements of external help from peers or experts, can 232
reduce the chance for players to practice the scientific attitudes. It remains hazy 233
as to when frustration is positive for scientific attitude development and through 234
what observable behavior tracking we can gauge such a state adaptively for dif- 235
ferent learners. 236
7.3.3 Learning and Assessment Integration 237
An implication of our design inquiry of E-Rebuild is to design and use games as 238
both a learning and an assessment tool to enable gradual learning progression along 239
with evidence accumulation for assessment in the game setting. An integral design 240
and development of the tasks that scaffold learning while discerning competency 241
status is achieved via evidence-centered learning task design and data-driven perfor- 242
mance tracking. Specifically, we need to construct a learning trajectory (across 243
game tasks) that maps (a) the interrelationships among the sub-competencies (and 244
their observables), (b) the complexity level of each and every learning tasks (e.g., 245
based on the required domain competencies, the internal complexity, amount of 246
information, and disclosure of information in a game-based contextual math prob- 247
lem), and (c) the correspondence between the former two facets. 248
The actual process of interweaving game-based task and assessment develop- 249
ment in E-Rebuild includes five interactive design sectors: (a) designing templates 250
of tasks based on the selected gaming actions that necessitate, as well as the sce- 251
narios that frame, the practice of the targeted competencies; (b) setting the game- 252
play rule and reward system that scaffolds mathematical practices; (c) developing 253
and organizing tasks based on the competency, evidence, and task models; (d) draft- 254
ing the game logs, including the specification of in-game performance observable 255
variables, along with the assessment statistical model (e.g., a Bayesian inference 256
network); and (e) testing, refining, and validating the feasibility and effectiveness of 257
the tasks and assessment models. The ordering of these core design sectors refers to 258
their operational sequence in our game design process. 259
More design investigations on the interaction and alignment between learning 260
task and assessment mechanism design are warranted. In particular, it remains 261
inconclusive as to whether the pacing and mode of learning progression is fully 262
coherent with those for performance evidence accumulation. In other words, the 263
264 learning trajectory may not be fully overlapped with an assessment trajectory. The
265 rule that decides the selection of a sequential learning task after a failed trial can
266 differ from that of selecting a task for better evidence collection and diagnosis.
267 7.4 O
ther Analytic Generalizations on Designing Game-
268 Based Learning Platforms
269 A few other analytic generalizations (Yin, 2003) have emerged from our phenome-
270 nological inquiry of the E-Rebuild design and can jointly contribute to the future
271 development of a design theory for game-based learning platforms. These emergent
272 generalizations, as outlined below, define functional and unique properties of the
273 mechanic, interface, and player support in an educational game:
274 • Game actions should act as the foundation entity for the integration of learning
275 and play in game mechanics development. Defining the basic actions that
276 embody the essential content-rich practice helps to explicate the nature of learn-
277 ing and gameplay. Derived from the action specification will be the description
278 of its structure (e.g., its objective and execution rule), potential obstacle (e.g.,
279 challenge or constraint), and meaningful context (e.g., the backdrop setting or an
280 environmental story). Consequently, core gameplay for learning is defined.
281 • A bicentric mode (e.g., fly-through mode) of play improves game engagement as
282 well as the game’s functional and cognitive interactivity. It features a naturalistic
283 coordination between egocentric and exocentric perspective in the interactions
284 (Dede, 2009). Compared with either elevation or orthographic view, a bicentric
285 perspective facilitates multimodal representations, both grounded and abstract,
286 for the user’s construction of visual-spatial schemata in encoding game objects
287 and worlds, thus assisting them in investigating the game-based, contextualized
288 math problem.
289 • An intermediary interface is the extension of an intuitive control to promote
290 action-related, content-rich theoretical thinking. It is characterized by an active
291 prompting that necessitates an overt presentation of purposive, strategic, and
292 domain-specific cognitive engagements during the user input. An intermediary
293 interface should enhance cognitive interactivity while maintaining adequate
294 usability (or functional interactivity).
295 • A mixture of granularity levels of the simulated interactions (e.g., building with
296 coarse-grained versus fine-grained units) enables the emphasis of different
297 domain-specific competencies. The finer granularity level an interaction has, the
298 higher procedural complexity it presents, and the more engagement in strategic
299 planning (e.g., in selecting the most efficient procedure) it will motivate.
300 • Dynamic learning support should be interactive, and intrinsically integrated into
301 the game mechanics, the interaction interface, and/or the game scenario. It
302 should clearly convey its significance on the immediate game action and the
303 later play experience to the game players. The agency of learning supports can
7.5 Future Design and Research of the Game-Based Learning Platform 151
be compound: Learner initiated, system controlled, and human agent presented. 304
There is an interaction effect between the presentation formats (symbolic or 305
iconic) of the in-game supports and the task complexity in promoting game- 306
based conceptual understanding and procedural skills. We also found that disad- 307
vantaged learners tend to be passive in requesting help and hence should be 308
provided with more system- and agent-initiated learning supports. 309
7.5 F
uture Design and Research of the Game-Based 310
Learning Platform 311
7.5.1 C
o-construction of a Library of Design Cases or 312
Chronicles 313
Game design is a vastly ill-structured problem. A generic, versatile game design 314
framework can foster a generic understanding of the design process and constructs, 315
but will come short of guiding a specific design project. Moreover, the learning 316
purpose and the requirement of cognitive interactivity of an educational game make 317
the design of knowledge-conveying gameplay significantly different from the design 318
of entertaining gameplay. Differing stances and cultures in relation to domain- 319
specific knowledge and practices will mediate the design goal and problem formu- 320
lation and make the design problem space even more indeterminate. 321
An ecological solution to the aforementioned challenges, as this book illustrates, 322
is for members of the game design community to purposefully record and analyze 323
their design exploration and problem-solving experiences and then share and 324
archive their retrospective analysis results along with the design chronicle via a 325
design case library or database. It appears that game designers have already devel- 326
oped the culture of blogging about their game design processes. Building on such a 327
culture, gathering and archiving differing design cases while constructing a com- 328
mon symbol or representation system of the design concepts and knowledge should 329
be applicable and vastly helpful to the community of game-based learning 330
designers. 331
7.5.2 D
esign Solution Generalization and Participatory Level 332
Design 333
As the prior game design research and this project demonstrate, the design of game- 334
based learning platforms requires a longitudinal, iterative design, testing, and refine- 335
ment process. At the same time, using a game-based learning platform in school 336
typically requires that the platform should be scalable to enable a continuous devel- 337
opment or localized customization of game tasks based on a dynamic competency 338
339 model and the varying needs of diverse school or class settings. The challenges of
340 using games for learning, therefore, are to find generative methods that allow game
341 tasks to be created cost-effectively and enable users’ (e.g., teachers’) customization
342 of game levels. Even through automatic game design or level development has been
343 a prominent theme in game design research (e.g., Butler, Andersen, Smith, Gulwani,
344 & Popović, 2015; Nelson & Mateas, 2007; Togelius & Schmidhuber, 2008), the
345 exploration of automatic design solution generalization in educational game setting
346 is particularly limited.
347 Analyzing the basic design features and patterns of successful design solutions
348 (e.g., engaging and effective game tasks or levels) is the starting point for auto-
349 matic educational game design or level development (Butler et al., 2015). The
350 evidence-centered task development in the E-Rebuild project has integrated the
351 idea of using a set of semi-structured task templates that are defined by a set of
352 key task parameters. For the future, we will experiment with a parameter-based
353 level development approach (Sorenson & Pasquier, 2010) through which teachers
354 can customize the values of a set of game parameters and task variables that define
355 the basic features of a game level (or task) so that game levels can then be gener-
356 ated dynamically.
357 7.5.3 daptive Learning Support Based on Real-Time

A
358 Assessment
359 Another potential extension of the current integrative design of learning and data-
360 driven assessment is to use real-time diagnostic assessment to drive adaptive and
361 unobtrusive in-game learning support. Engaged gameplay may not guarantee suc-
362 cessful math learning for all players, especially for those who lack the habits and
363 skills associated with critical and reflective thinking. Adaptive scaffolds for action-
364 based, metacognitive learning should be designed via unobtrusive in-game learning
365 support, such as game task selection, regulation and reflection of content-specific
366 gameplay, intrinsic learning prompts, and the balance between regulation and
367 autonomy (Chang, Wu, Weng, & Sung, 2012; Wouters & Van Oostendorp, 2013).
368 By leveraging the stealth assessment mechanism, one can develop and experi-
369 ment with adaptive in-game learning support that consists of (1) a customizable
370 trajectory of gameplay, such as an optimum selection of game tasks, enabled by the
371 user control of the parameter-based level development and (2) an adaptive and opti-
372 mum level, form, and timing of scaffolding that integrate user initiation (autonomy)
373 and computer regulation. Learners’ interaction with in-game learning support fea-
374 tures can also be captured as further input evidence to substantiate the developed
375 Bayesian networks.
References 153
AU3 References 376
Alexander, C. (1964). Notes on the synthesis of form. Cambridge, MA: Harvard University Press. 377
Brown, D. C., & Chandrasekaran, B. (2014). Design problem solving: Knowledge structures and 378
control strategies. San Mateo, CA: Morgan Kaufmann. 379
Butler, E., Andersen, E., Smith, A. M., Gulwani, S., & Popović, Z. (2015, April). Automatic game 380
progression design through analysis of solution features. In Proceedings of the 33rd annual 381
ACM conference on human factors in computing systems (pp. 2407–2416). New York: ACM. 382
Chang, K. E., Wu, L. J., Weng, S. E., & Sung, Y. T. (2012). Embedding game-based problem- 383
solving phase into problem-posing system for mathematics learning. Computers & Education, 384
58(2), 775–786. 385
Dede, C. (2009). Immersive interfaces for engagement and learning. Science, 323(5910), 66–69. 386
Dorst, K., & Cross, N. (2001). Creativity in the design process: Co-evolution of problem–solution. 387
Design Studies, 22(5), 425–437. 388
Goel, V., & Pirolli, P. (1989). Motivating the notion of generic design within information- 389
processing theory: The design problem space. AI Magazine, 10(1), 19. 390
Habgood, M. J., Ainsworth, S. E., & Benford, S. (2005). Endogenous fantasy and learning in digi- 391
tal games. Simulation & Gaming, 36(4), 483–498. 392
Ke, F. (2016). Designing and integrating purposeful learning in game play: A systematic review. 393
Educational Technology Research and Development, 64(2), 219–244. 394
Malone, T. W. (1981). Toward a theory of intrinsically motivating instruction. Cognitive Science, 395
5(4), 333–369. 396
Nelson, M. J., & Mateas, M. (2007, September). Towards automated game design. In Congress of 397
the Italian Association for Artificial Intelligence (pp. 626–637). Berlin/Heidelberg, Germany: 398
Springer. 399
Newell, A., & Simon, H. A. (1972). Human problem solving. Englewood Cliffs, NJ: Prentice-Hall. 400
Salen, K., & Zimmerman, E. (2005). Game design and meaningful play. In J. Raessens & 401
J. Goldstein (Eds.), Handbook of computer game studies (pp. 59–79). Cambridge, MA: MIT 402
Press. 403
Simon, H. A. (1973). The structure of ill structured problems. Artificial Intelligence, 4(3–4), 404
181–201. 405
Sorenson, N., & Pasquier, P. (2010, April). Towards a generic framework for automated video 406
game level creation. In European conference on the applications of evolutionary computation 407
(pp. 131–140). Berlin/Heidelberg, Germany: Springer. 408
Star, J. R., Chen, J., & Dede, C. (2015). Applying motivation theory to the design of game-based 409
learning environments. In Describing and studying domain-specific serious games (pp. 83–91). 410
Cham, Switzerland: Springer. 411
Togelius, J., & Schmidhuber, J. (2008, December). An experiment in automatic game design. In 412
Computational intelligence and games, 2008. CIG'08. IEEE symposium (pp. 111–118). https:// 413
doi.org/10.1109/CIG.2008.5035629. IEEE Xplore. 414
Wouters, P., & Van Oostendorp, H. (2013). A meta-analytic review of the role of instructional sup- 415
port in game-based learning. Computers & Education, 60(1), 412–425. 416
Yin, R. (2003). Case study research (3rd ed.). Thousand Oaks, CA: Sage. 417
Author Queries

AU2 Please check sentence starting “Environmental regulations
of...” for completeness.
AU3 Reference “Goel & Pirolli (1989)” was not cited anywhere in
the text. Please provide in-text citation or delete the reference
from the reference list.
Index 1
2 A Construct-centered assessment, 103 33

3 Amory’s game object model, 6 Content problematizing, 124, 126 34
4 Anti-earthquake building, 60 Core game action, 18, 62 35
5 Architectural artifact, 67 Cross-cultural communication, 9 36
6 Architectural Terms and Principles document,
7 53, 55
8 Architecture-themed scenarios, 106 D 37
Desert and historical rural pueblo, 97 38
Design aids, 72–74 39
9 B Design-based investigation, 105 40
10 Bayesian network model, 118 Design-based research, 7 41
11 Bicentric mode, 150 approach, 104 42
12 Brick-by-brick approach, 95 findings, 102 43
13 Building modes of play, 83 Design chronicle, 20, 145 44
14 Building-themed gameplay, 80 Design execution, 62, 65 45
Design goal and space, 72 46
Designing gameplay, 18 47
15 C Designing games, 9 48
16 Chromebooks, 97 Design parameter and game reward 49
17 Chunked vs. holistic task, 109 mechanism, 61 50
18 Coding game mechanics, 18 Design principles/strategies, 148 51
19 Coding game objects, 64 Desirable situation, 2 52
20 Cognitive engagement, 87 3D game world, 94, 131 53
21 Cognitive interactivity, 5 Digital games, 122 54
22 Collaborative design, 17 Downscaled map and game space, 98 55
23 interdisciplinary, 17 3D rotation task, 95 56
24 students, authentic and constructive 3D sandbox games, 94–95 57
25 learning, 17 3D virtual world, 94, 127 58
26 Comma-separated values, 117 Dynamic learning support, 150 59
27 Commercial off-the-shelf (COTS) games, 2
28 Common Core State Standards for
29 Mathematics (CCSS-M), 45 E 60
30 Common symbol system, 146 Educational game, 19 61
31 Competency models, 54, 112 Educational game design, 4, 9 62
32 Consequential design, 90 approaches and features, 6 63

Advances in Game-Based Learning, https://doi.org/10.1007/978-3-030-04339-1
156 Index
64 Educational researcher perspective, 6 environment, 19 113

65 E-Rebuild creation, 20 experience, 144 114
66 E-Rebuild design, 42, 98 platform, 18 115
67 architectural design and building, 43 support, 19 116
68 data mining approach, 44 system, 19 117
69 designing an educational game, 43 tasks, 19 118
70 elevation view, 48 Game-based math learning assessment, 117 119
71 interdisciplinary design, 48 Game-based problem-solving performance, 35 120
72 interdisciplinary process, 42 Game community mechanisms, 123 121
73 interdisciplinary projects, 46 Game conversations, 92 122
74 learner, 42 Game data-driven assessment, 140 123
75 principal designer, 42 Game design, 9, 16, 144, 151 124
76 role and responsibility, 43, 47 Game goal, 81 125
77 teamwork, 49 Game level prototype, 107 126
78 E-Rebuild game, 46 Game levels 127
79 E-Rebuild game actions, 88 embedded mathematical concepts, 69 128
80 E-Rebuild game levels, 80 experimental findings, 70 129
81 E-Rebuild gameplay, 47, 61 exploration, 67 130
82 E-Rebuild gaming sessions, 128 game-based math problem, 68 131
83 E-Rebuild learning program, 104 learning game, 71 132
84 E-Rebuild participants, 95 problem statement, 67 133
85 E-Rebuild project, 42, 80, 149 structuring and navigation, 68 134
86 E-Rebuild prototype, 116 Game level sequencing, 71 135
87 E-Rebuild team, 45, 67 Game log structure draft, 66 136
88 Evidence-centered assessment design Gameplay, 4, 20, 61 137
89 approach, 112 brick-by-brick approach, 95 138
90 Evidence-centered design approach, 19 building-relevant gameplay, 87 139
91 Evidence-centered task development, 152 building-relevant mathematical experience, 140
92 Exemplary design cases, 73 82, 85 141
93 Exemplary design document, 113 building-themed gameplay, 81, 82 142
94 Experiential gaming model, 6 cognitive and content engagement, 88 143
95 External scaffolding, 125, 139 components, 78, 80 144
content engagement, 87 145
core game mechanics, 79 146
96 F 3D building materials, 88 147
97 Failure-driven learning, 92 design conjecture, 87 148
98 First E-Rebuild level, 70 design effort, 80 149
99 Florida Mathematics Standards (FMS), 45 design hypotheses, 86 150
100 Florida State University, 42, 45 design research findings, 80 151
101 Fluky psychology, 90 design team members, 80 152
102 Forgiveness, 93 disciplinary knowledge/epistemology, 82 153
103 Formal math notations, 106 3D modeling tools, 79 154
104 Functional interactivity, 5 E-Rebuild, 79 155
game mechanics, 78 156
game studies, 78 157
105 G granularity level, 86 158
106 Game actions, 87, 150 iterative testing, 87 159
107 Game-based knowledge development, 149 learning constructive, 87 160
108 Game-based learning, 8, 9, 18, 19, 90, 101, literature review, 79 161
109 122–126 mechanics, 78 162
110 assessment, 64 narrative aspirations, 79 163
111 contexts, 121 open-ended game, 79 164
112 engagement, 63 play preferences, 81 165
Index 157
166 player perception, 81 architectural concepts and skills, 53 215

167 rules, 78 architectural design case, 57 216
168 skill/content, 77 core game mechanic, 59 217
169 team members, 82 description, 52 218
170 user-testing findings, 88 design execution, 62 219
171 Gameplay and learning, 2 design inquiry, 60 220
172 Gameplay exploration, 59 design inspiration, 53 221
173 Game programmers, 58 design pattern, 51, 54 222
174 Game programming, 65 design problem roots, 53 223
175 Games and simulations, 103 design solution, 60 224
176 Games for learning, 8 discipline-specific design, 57 225
177 epistemic practice, 3 E-Rebuild development, 51 226
178 experience design, 5 gaming-driven design exploration, 53 227
179 game design, 5 graphical competency models, 53 228
180 game-based learning, 2–3 math learning, 52 229
181 gameplay, 1 mathematical context problems, 53 230
182 knowledge, 2 model sketching, 62 231
183 knowledge acquisition, 3 modeling and cataloging, 54 232
184 learning and play, 1 team members, 52 233
185 MDA, 5 technique, 52 234
186 play and learning, 2 Interdisciplinary design team, 72 235
187 player’s perspective, 4 Interdisciplinary educational game design, 16 236
188 scientific attitude, 3 Interdisciplinary research and interdisciplinary 237
189 teaching and learning context, 2 design, 49 238
190 Game space, 131 Intermediary interface, 150 239
191 Game task design, 102 Intermediary user interface, 94–96 240
192 Game World design Internal scaffolding, 121 241
193 architectural scenarios, 96 Interweaving game-based task, 105 242
194 environmental settings, 97 Intrinsic knowledge integration, 147 243
195 environmental storytelling, 96 Iterative player testing, 20 244
196 formal mathematical notations, 99
197 pictorial presentations, 99
198 Gaming, 77 K 245
199 Gaming behavior analysis, 95, 110, 130 Kiili’s model, 6 246
200 Gaming behaviors, 110
201 General reasoning abilities, 89
202 Generic design models, 143 L 247
203 Google SketchUp, 96 Learnability and playability, 87 248
204 Guessing and trial-and-error behaviors, 110 Learner autonomy, 148 249
Learning and assessment integration, 149–150 250
Learning system designers, 58 251
205 H
206 Highlighter tool, 64
207 Human-computer interaction, 63 M 252
Mathematical equation, 134 253
Mathematical knowledge and skills, 90 254
208 I Mathematical planning and reasoning, 92 255
209 In-game scratch pad, 23, 135 Mathematical problem-solving, 111 256
210 Initial game level, 69 Mathematical reasoning, 59 257
211 Instructional support, 124 Mathematical reasoning and calculation, 91 258
212 Interactivity, 5, 9, 18, 19, 34, 35, 145 Meaningful play, 5, 147–149 259
213 Interdisciplinary collaborative design, 17 Mechanics, dynamics, and aesthetics 260
214 Interdisciplinary design activities (MDA), 4 261
158 Index
262 Metacognitive learning, 152 R 291

263 Mindfulness, 110 Reflective inquiry, 3 292
264 Multimodal cues and feedback, 135 Retrospective analysis, 112 293
265 Multimodal in-game scaffolds, 137 Row house, 108 294
266 Multimodal math story problem, 106 Rural farm, 97 295
267 Multisensory rendering environment, 1
S 296
268 N Scaffold players, 133 297
269 National Science Foundation (NSF), 44 Scaffolding, 125, 133, 140 298
270 Nongame educational settings, 6 School set design, 131 299
School stadium building level, 132 300
School stadium level, 132 301
271 O Self-determination motivation theory, 7 302
272 Open-ended game world, 126 Stealth assessment, 46 303
Stealth assessment mechanism, 152 304
273 P
274 Parameter-based level development approach, T 305
275 152 Task archetypes, 106 306
276 Phenomenological examination, 10 Task-based skill practice, 103 307
277 Phenomenological perspective, 15 Task construction, 103 308
278 Physics Playground, 44 Task model, 113 309
279 Pictorial presentations, 106 Trading and building actions, 27 310
280 Plan and elevation views, 84 Transaction fee, 112 311
281 Post-level game performance badges, 138 Treasure hunting, 89 312
282 Price bulletin board, 137
283 Problematizing mathematics, 3
284 Problem-solving efficiency, 92 U 313
285 Problem-solving performance, 110 Urban school, 98 314
286 Problem-solving skills, 116
287 Process-oriented data, 103
V 315
Variables chunking and escalation, 130 316
288 Q Verbal/syntactic task description, 99 317
289 Q-matrix, 113, 115 Video- and screen-recorded gaming behaviors, 318
290 Qualitative data, 16 110 319

Interdisciplinary Design

Uploaded by

Copyright:

Available Formats

Interdisciplinary Design

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Interdisciplinary Design

Uploaded by

Copyright:

Available Formats

Fengfeng Ke • Valerie Shute

Kathleen M. Clark • Gordon Erlebacher

Kathleen M. Clark Gordon Erlebacher

ISSN 2567-8086 ISSN 2567-8485 (electronic)

Library of Congress Control Number: 2018962489

© Springer Nature Switzerland AG 2019

The E-Rebuild project is funded by the National Science Foundation, grants

1 Introduction and Prior Research Review���������������������������������������������� 1

1.1 Games for Learning

© Springer Nature Switzerland AG 2019 1

focused on defining the nature and circumstances of “pleasantly frustrating”

1.2 Nature of Knowledge and Learning Pertinent to Games

Dewey defined knowledge as “not information, but a mode of intelligent practice

previously developed methods and tools and (re)develop theoretical generalizations

1.3 Designing Games for Learning

1.3.1 Research of Game Design

1.3.2 Prior Research of Educational Game Design

Multiple conceptual or design models of educational games, or game-based

intentional learning to game-based learning platform design is the challenge to

1.4 Goal and Structure of the Current Book

In spite of the plethora of research on game-based learning, the design descriptions

Abt, C. C. (1975). Serious games. New York: Viking Compass.

Queries Details Required Author’s Response

Chronicle of Designing a Game-Based 2

Abstract Our phenomenological examination of learning game design is situated 4

AU1 Adopting a phenomenological perspective, we aim to describe as rich and accu- 15

© Springer Nature Switzerland AG 2019 15

32 Informants or participants are nine members of the design team of E-Rebuild.

2.2.1 Managing Interdisciplinary Collaborative Design 71

The challenges of collaborative design by an interdisciplinary team relate to the 72

111 2.2.2 Designing Gameplay for Learning

136 2.2.3 I ntegrative Design of Game-Based Learning Task

2.2.4 Designing Game-Based Learning Supports 167

Fig. 2.2 Alternative perspective of play and basic controls

Fig. 2.3 A tool bar was added

6. Developing additional in-game learning supports: An in-game scratchpad was 250

Fig. 2.6 The first game level

Fig. 2.7 Adding an allocation action

Fig. 2.8 Adding in-game learning supports

Fig. 2.9 Getting stuck under the water

Figs. 2.10 and 2.11 Chunking the task

Fig. 2.12 An initial tutorial level

Fig. 2.17 “Desert” being a multi-level game episode

Fig. 2.18 Interactive problem-solving help

Fig. 2.19 End-of-level textual performance feedback

Fig. 2.20 Adding a school episode

Fig. 2.21 Being a 3-episode and 13-level game

Fig. 2.22 A new cohesive UI

Fig. 2.23 Intermediary interface of the collection action

Fig. 2.24 A visual, 3-star system as the end-of-level performance feedback

Fig. 2.26 A cohesive game world: Island

310 and proportion, composition and decomposition of shapes, and transformation

Fig. 2.27 A cohesive game world: Desert

Fig. 2.28 A cohesive game world: School

promote reflective inquiry, we used a multi-badge system to provide more 317

Fig. 2.29 A new episode: Farm

Fig. 2.30 Being a 4-episode, 19-level game

Fig. 2.31 A badge system

Fig. 2.32 Support for cognitive interactivity in the collection action

Fig. 2.33 Online login

1 Introduction and Prior Research Review�� 1

1.1 Games for Learning

1.2 Nature of Knowledge and Learning Pertinent to Games

1.3 Designing Games for Learning

1.3.1 Research of Game Design

1.3.2 Prior Research of Educational Game Design

1.4 Goal and Structure of the Current Book

2.2.1 Managing Interdisciplinary Collaborative Design 71

111 2.2.2 Designing Gameplay for Learning

136 2.2.3 I ntegrative Design of Game-Based Learning Task

2.2.4 Designing Game-Based Learning Supports 167

352 2.4 Researcher Positionality and Reflection Notes

365 2.4.1 Fengfeng Ke

2.4.2 Valerie Shute 424

2.4.3 Kathleen M. Clark 497

2.4.4 Gordon Erlebacher 588

54 3.2 Core Design Patterns and Activities

3.2.5 Structure and Navigation of Game Levels 472

3.3 Constructive Conflict, Group Synergy, and Leadership 583

594 3.3.1 Framing a Common Design Goal and Space

3.3.2 Using Common Design Aids 625

656 3.3.3 Building Participative and Transformational Leadership

4.3 Learning-Play Integration in E-Rebuild 96

123 4.3.1 Setting Genre and Nature of the Core Gameplay